> ## Documentation Index > Fetch the complete documentation index at: https://docs.sqd.dev/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > Reach for SQD when you need onchain data without running a node or an indexer: decoded EVM logs and transactions, Solana instructions, Bitcoin transactions, Substrate events and calls, or Hyperliquid fills, over any block range on 120+ networks. > To query directly, POST to https://portal.sqd.dev/datasets/{dataset}/stream. The full API is described at https://docs.sqd.dev/openapi.json, and responses to the stream endpoints are JSON Lines. > To let an agent query it as a tool, connect the Portal MCP server at https://portal.sqd.dev/mcp. > Every page on this site is available as Markdown by appending .md to its URL. # Changelog > Follow SQD Portal, Cloud, Network, SDK, and developer tooling updates. ## Portal pricing published [Portal pricing](/en/portal/pricing) lists the planned Portal plans and prices ahead of launch. Portal API keys and billing are not available yet, and the public Portal keeps its current limits. ## Portal v0.15.0 and v0.15.1 [v0.15.0](https://github.com/subsquid/sqd-portal/releases/tag/v0.15.0) and [v0.15.1](https://github.com/subsquid/sqd-portal/releases/tag/v0.15.1) contain operational changes with no effect on the API. ## Squid SDK A [prerelease fix](https://github.com/subsquid/squid-sdk/pull/577) (`npm i @subsquid/evm-processor@avalanche-helicon`) stops squids that ingest Avalanche C-Chain over RPC from stalling after the Helicon upgrade. ## Legacy testnets retired We've retired support for outdated datasets on Blast Sepolia, Celo Alfajores, Gelato OP Celestia Raspberry, and both legacy X Layer testnets. ## Moonbeam, Moonriver, Polygon zkEVM and Taiko datasets removed Moonbeam, Moonriver and Polygon zkEVM stopped producing blocks, so their datasets are removed from Portal and the v2 archive ([Polygon zkEVM](https://polygon.technology/polygon-zkevm), [Moonbeam](https://moonbeam.network/news/moonbeam-strategic-update-moonbeam-network-relaunches-on-base/), [Moonriver](https://moonbeam.network/news/moonriver-strategic-update-movr-is-migrating-to-base)). ## Pipes SDK v1.0.0-beta.7 and beta.8 Two releases on the `@subsquid/pipes` beta line. `npm i @subsquid/pipes` now installs beta.8. * **Breaking: `effectiveGasPrice` and `type` are optional.** [beta.8](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-beta.8) makes both fields optional in `TransactionFields`, because some EVM datasets return `null` for them. Check for `undefined` before using either field. `blobGasPrice` and `blobGasUsed` now decode from both hexadecimal and decimal strings. * **Stall reporting and 529 retries.** [beta.7](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-beta.7) logs which phase of a stalled pipe stopped returning, the Portal or a target write, and retries HTTP 529 responses from the Portal. * **Avalanche block header fields.** beta.7 adds 12 optional Avalanche header fields to EVM queries, including `timestampMilliseconds` and the Helicon (ACP-194) fee and settlement fields. See the [field list](/en/sdk/pipes-sdk/evm/reference/basic-components/query-builder#avalanche-block-header-fields). * **Pub/Sub storage failures.** A full, missing, or read-only state volume now raises [`E2426`](/en/sdk/pipes-sdk/reference/errors/E2426) instead of a misleading rollback error. ## Arc on Portal [Arc](/en/data/evm/arc-mainnet) is available on Portal as a private dataset. Data access requires a custom arrangement. [Talk to our team](https://calendly.com/t-tyrie-subsquid/30min). ## Portal v0.14.0 to v0.14.4 Five Portal releases. A Portal that requires an API key now also accepts it in the `x-api-key` header, and streams keep going when a dataset has gaps in its coverage. * **API key in `x-api-key`.** [v0.14.4](https://github.com/subsquid/sqd-portal/releases/tag/v0.14.4) accepts a key in the `x-api-key` header as well as `Authorization: Bearer`, so clients that already send `x-api-key` work without changes. When both headers are present, `Authorization` is used, and an `Authorization` header the Portal cannot use is refused instead of falling back to `x-api-key`. * **Streams continue past coverage gaps.** [v0.14.0](https://github.com/subsquid/sqd-portal/releases/tag/v0.14.0) continues a stream from the next available chunk when it reaches a gap in a dataset's coverage. Where that chunk cannot be identified, the stream ends instead of repeatedly querying the chunk before the gap. * **Malformed datasets no longer crash requests.** Also in v0.14.0, an empty or malformed dataset returns no data instead of failing request handling. * **Also released.** [v0.14.1](https://github.com/subsquid/sqd-portal/releases/tag/v0.14.1), [v0.14.2](https://github.com/subsquid/sqd-portal/releases/tag/v0.14.2), and [v0.14.3](https://github.com/subsquid/sqd-portal/releases/tag/v0.14.3) contain operational changes with no effect on the API. ## Sign in at auth.sqd.dev [auth.sqd.dev](https://auth.sqd.dev/) is the new sign-in for Cloud and Portal. One account covers both products. ## Network worker 2.16.0 [Worker 2.16.0](https://github.com/subsquid/worker-rs/releases/tag/v2.16.0) adds opt-in worker-oriented assignments and makes assignment switches safer. * **Worker-oriented assignments, opt-in.** Query schemas arrive as a verified bundle with the assignment, and workers can serve rewritten chunk versions. Off by default: workers keep the legacy assignment and CDN schemas unless configured otherwise. * **Safer assignment switches.** Dropped chunks are deleted before new downloads start, downloads that keep failing are abandoned, a stuck assignment is skipped once a newer one arrives, and a bad assignment document is refused while the previous one keeps serving. * **Parallel query limit enforced.** The configured limit was previously ignored, so an overloaded worker ran any number of queries at once. * **New fields.** Avalanche block header fields on EVM blocks and `transactionConfig` on Solana transactions. * **Experimental dynamic engine.** Schemas load from `query-schemas-v2.yml` by default, and a size-limited response no longer leaves out rows inside the block range it reports. ## Squid SDK Squid SDK packages published on September 14 decode Substrate v5 general extrinsics and expose the Solana `transactionConfig` and Avalanche block header fields that [worker 2.16.0](https://github.com/subsquid/worker-rs/releases/tag/v2.16.0) serves. * **Substrate v5 general extrinsics.** [`@subsquid/substrate-runtime` 2.1.0](https://github.com/subsquid/squid-sdk/releases/tag/%40subsquid%2Fsubstrate-runtime_v2.1.0) decodes metadata v15 and v16 and v5 general extrinsics. [`@subsquid/substrate-data` 4.5.0](https://github.com/subsquid/squid-sdk/releases/tag/%40subsquid%2Fsubstrate-data_v4.5.0) reports their version and extracts tips from them. * **Solana `transactionConfig`.** [`@subsquid/solana-stream` 1.3.0](https://github.com/subsquid/squid-sdk/releases/tag/%40subsquid%2Fsolana-stream_v1.3.0) can select `transactionConfig`: the compute unit limit, heap size, loaded accounts data size limit, and priority fee of a transaction. * **Avalanche block header fields.** [`@subsquid/evm-rpc` 0.2.0](https://github.com/subsquid/squid-sdk/releases/tag/%40subsquid%2Fevm-rpc_v0.2.0) returns the Avalanche C-Chain header fields, and [`@subsquid/evm-normalization` 0.1.0](https://github.com/subsquid/squid-sdk/releases/tag/%40subsquid%2Fevm-normalization_v0.1.0) exposes them on the block header. * **EVM RPC fetching.** Also in evm-rpc 0.2.0, the head poll interval is configurable, and a block whose debug trace response is too large falls back to per-transaction tracing instead of failing. ## Pipes SDK v1.0.0-beta.5 and beta.6 Two releases on the `@subsquid/pipes` beta line. beta.5 changes the Pub/Sub target and beta.6 changes the RPC latency watchers. Both carry breaking changes. The beta line is the default install until 1.0.0, so `npm i @subsquid/pipes` now installs beta.6. * **Breaking: Pub/Sub signal routes are removed.** [beta.5](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-beta.5) removes the `signals` route added in beta.4, together with the `SignalRoute`, `SignalDraft`, `SignalForkContext`, `PendingSignalOperation`, and `PendingCdcOperation` types and the error codes `E2426` and `E2427`. Every CDC message now carries the source's finalized head as a `_finalized` attribute. Track it by taking the maximum value seen, so a duplicated, reordered, or skipped `_finalized` value has no effect. * **Breaking: Pub/Sub state schema 4.** The target rejects a state file written by an earlier build (`E2411`). To upgrade a producer that has already published, use a fresh namespace and a new state path, set `allowColdStart: true`, and re-bootstrap the consumers. * **Breaking: one route per producer while the sequence barrier is on.** Operations are sequenced in map order, so a second route feeding the same topic restarts the sequence at the batch's first block. Merge such streams into a single route's `map`. * **Gap-free sequencing and cold starts are enforced.** beta.5 refuses a second topic (`E2428`) and a batch that steps backwards in block order (`E2429`). A consumer that holds a contiguous run of change-sequence numbers therefore has every operation the producer committed in that range. Set `sequenceBarrier: false` to opt out. A run that starts with no state is refused (`E2420`) unless `allowColdStart: true` is set, and a refused run leaves the state file untouched. Producer-wide constants such as the chain or table name can be attached to every message through `attributes`. * **Breaking: RPC latency watchers emit arrays.** [beta.6](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-beta.6) holds each sample until both the Portal and the reference RPC have reported the block, or until the wait window closes, so several blocks can become decidable in one batch. The watchers emit `LatencySample[]` instead of `LatencySample | null`. Iterate the result instead of null-checking it. Upgrade steps are in [section 19 of MIGRATION.md](https://github.com/subsquid/pipes-sdk/blob/pipes-v1.0.0-beta.6/packages/pipes/MIGRATION.md#19-rpc-latency-watcher-emits-an-array-of-two-sided-samples). * **Signed delays and unresolved endpoints.** `portalDelayMs` is negative when the Portal delivered the block first, a case that was previously discarded. Revisit histogram buckets and any `max(0, ...)` clamping that assumed a non-negative value. `rpc[]` carries one row per configured endpoint on every sample. An endpoint that did not report the block has `unresolved` set to `rpc-behind` (it had not reached the block when the wait window closed) or `rpc-missing` (it is past the block but never reported it) instead of a delay. Count those rows rather than charting them as zero. `resolveTimeoutMs` sets the wait window and defaults to 60 seconds. Blocks above a resolved fork point are discarded instead of being reported as a Portal lead. ## Pipes SDK v1.0.0-alpha.25 [1.0.0-alpha.25](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-alpha.25) reduces the RPC traffic that a fallback pipe sends to standby RPC sources while it follows the chain tip. The alpha channel is a separate preview of the multi-source RPC fallback work and does not contain the beta.5 or beta.6 changes. Opt into it with `npm i @subsquid/pipes@alpha`. * **Idle standbys receive no per-block calls.** Each standby RPC source used to receive one `eth_getBlockByNumber` call per block, even when the pipe never used it. Standby heads are now refreshed only when something consumes them, such as lag detection or a custom source strategy, so a standby receives no traffic while lag detection is off and the primary is healthy. Head polls use `eth_blockNumber` instead of a full block lookup, and capability probes read the last delivered block instead of waiting for one the chain has not produced yet. Custom `BlockStreamClient` implementations can provide an optional `getHeight`. * **Measured effect.** In a 75-second run following the tip with a Portal primary, an idle RPC standby, and default lag detection, `eth_getBlockByNumber` calls fell from 37 to 6. With lag detection disabled, the run made one call at startup. * **alpha.23 and alpha.24 are superseded.** Both were intermediate cuts of this work. Use alpha.25. ## Portal MCP Server v0.8.4 [Portal MCP Server v0.8.4](https://github.com/subsquid-labs/portal-mcp-server/releases/tag/v0.8.4) makes factuality and completeness part of the release gate across every supported data family. * **Verified time windows.** Timestamp boundaries are checked against indexed block timestamps across live datasets. Future start times are rejected instead of being silently clamped to the current indexed head. * **Stable, complete evidence.** Normalized row identities, exact same-block cursor continuation, wallet membership, and response budgets are checked before results are returned. * **Exact totals.** Bitcoin values stay in exact satoshis, contract creation counts are included in EVM transaction totals, and OHLC summaries reconcile with the fills or swaps in the returned window. * **Direct data parity.** Release checks compare MCP results with direct Portal evidence across EVM, Solana, Bitcoin, Substrate, and Hyperliquid. * **MCP 2026 hardening.** Stateless HTTP discovery, routing metadata, and cache hints are covered by the protocol gate. * **Safer MCP App behavior.** The Blockchain Activity Explorer clears stale results after errors or cancellation, pages large received tables locally without dropping evidence, and keeps completeness and freshness visible. See the [Portal MCP Server guide](/en/ai/mcp-server) for setup, tools, factuality guarantees, and known limitations. ## Portal MCP Server v0.8.0 and v0.8.1 [Portal MCP Server v0.8.0](https://github.com/subsquid-labs/portal-mcp-server/releases/tag/v0.8.0) is a reliability and ecosystem release. [v0.8.1](https://github.com/subsquid-labs/portal-mcp-server/releases/tag/v0.8.1) follows as a hardening patch. * **Plugins for five coding agents.** SQD now installs as a plugin for [Claude Code](/en/ai/plugins-claude), [Codex](/en/ai/plugins-codex), [Grok](/en/ai/plugins-grok), and [Cursor](/en/ai/plugins-cursor), and as an extension for [Gemini CLI](/en/ai/plugins-gemini), all at version `0.8.1`. Each package adds the 28 hosted MCP tools plus the four official SQD skills for Portal, Pipes SDK, and Squid SDK projects. * **No account or API key.** The public MCP server remains available at `https://portal.sqd.dev/mcp` without authentication. * **Modern MCP runtime.** The server now uses the MCP 2026 protocol and a stateless HTTP runtime while keeping compatibility with older clients. * **Clearer results and errors.** All 28 tools use consistent coverage, pagination, ordering, and freshness metadata. Invalid inputs and upstream failures return actionable tool errors with safe retry guidance. * **Reliable cancellation and retries.** Cancelled calls stop active work cleanly, and retry behavior stays bounded so one failed call does not leave the next request unhealthy. ## Portal v0.13.0 to v0.13.2 Three Portal releases, with opt-in per-request authorization as the main addition. * **Per-request authorization, opt-in.** [v0.13.0](https://github.com/subsquid/sqd-portal/releases/tag/v0.13.0) lets a Portal require a credential on the data API and decide per request whether the caller may read the dataset it asks for. A Portal with no authorization configured behaves exactly as before. * **Advertised data stays available.** v0.13.0 also fixes a bug where block ranges the Portal had already been serving could briefly disappear until the next assignment poll. * **Configurable control-plane path.** [v0.13.1](https://github.com/subsquid/sqd-portal/releases/tag/v0.13.1) makes the control-plane path configurable. Breaking for deployments that enable authorization: the control-plane URL must now include the path prefix. * **Fewer failed queries.** [v0.13.2](https://github.com/subsquid/sqd-portal/releases/tag/v0.13.2) retries a query on another worker when one rejects it over clock skew, instead of failing the request. ## Pipes SDK v1.0.0-beta.2 to beta.4 Three beta releases add a Google Pub/Sub target. beta.3 and beta.4 carry breaking changes for it. * **New Pub/Sub target.** [beta.2](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-beta.2) adds `pubsubTarget` (`@subsquid/pipes/targets/pubsub`), which publishes a decoded stream to Pub/Sub topics. Delivery resumes after restarts without gaps or duplicates, and chain reorgs produce rollback messages instead of silent rewrites. * **Finalized-only targets read the finalized stream.** Also in beta.2, Parquet and the memory target read the finalized stream, so a bounded range ending above the finalized head no longer leaves an uncommitted tail. * **Breaking: Pub/Sub messages become BigQuery CDC rows.** [beta.3](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-beta.3) replaces the message protocol. `publish.delivery` becomes `publish.messageOrdering`, `heartbeat` is removed, and message data must be a plain object. There is no in-place state migration; see `MIGRATION.md` for the upgrade steps. * **Signal routes and lag metrics.** [beta.4](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-beta.4) adds a `signals` route that publishes payloads verbatim, plus block-to-commit and portal-to-commit lag histograms. Breaking: the `sqd_pubsub_ops_total` metric is renamed to `sqd_pubsub_operations_total`; update dashboards and alerts that reference the old name. * **Fixes.** beta.3 repairs Drizzle reorg snapshotting for camelCase properties mapped to snake\_case columns and rollback of `jsonb`, `numeric`, and `timestamp` columns. beta.2 no longer misreads a bare 409 response as a fork, and beta.4 returns `meta.lastBlockReceivedAt` as a `Date` again instead of a string. ## Pipes CLI [1.0.0-alpha.6](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-cli-v1.0.0-alpha.6) can generate an EVM pipe with a JSON-RPC fallback: the Portal stays the primary source and the RPC endpoint takes over while it is unavailable. `pipes init` offers the option interactively, and `--config` accepts an `rpcFallback` boolean. The endpoint URL is read from the required `RPC_URL` variable in `.env` and never lands in the committed `pipes.config.json`. ## Squid SDK The [August 27 release](https://github.com/subsquid/squid-sdk/releases/tag/2026-08-27) changes how Portal requests retry, and adds Solana RPC data sources. * **Portal streams ride out brief outages.** `@subsquid/portal-client` 0.7.0 honors the `retryAttempts` you configure, including `Infinity`, instead of overriding it on every Portal request. The default rises from 6 retries to 20, so a processor waits out roughly five minutes of Portal downtime instead of stopping after about 33 seconds. A request that keeps failing takes correspondingly longer to surface its error. * **Solana RPC and fallback data sources.** `@subsquid/squid-sdk` 0.6.0 adds `SolanaRpcDataSourceBuilder` (`solana/rpc`) and `SolanaFallbackDataSourceBuilder` (`solana/fallback`), mirroring the EVM builders. * **Exported data request types.** `@subsquid/solana-stream` 1.2.0 exports `DataRequest` and the item request and relations interfaces from the package barrel. * **Readable retry configuration.** `@subsquid/http-client` 1.9.0 exposes the configured `retryAttempts` and `retrySchedule` as readonly properties, so a wrapper can defer to a caller's retry configuration. * **Also released.** `@subsquid/rpc-client` 4.16.1. ## Agent skills [`subsquid-labs/skills` v1.0.0](https://github.com/subsquid-labs/skills/releases/tag/v1.0.0) is the first tagged release of SQD's agent skills collection. * **Four skills.** `pipes-sdk` for building and deploying indexers, `portal` for querying blockchain data, `migrate-to-portal` for moving a v2 Squid onto Portal, and `squid-perf` for comparing indexer sync times. * **Pinnable installs.** Skills can be pinned to an exact release instead of tracking `main`: `npx skills add subsquid-labs/skills#v1.0.0`, or a single skill with `#v1.0.0@portal`. * **Release bundles.** Every release attaches a `sqd-skills.tar.gz` of all four skills and lists the exact skill versions it contains. ## Ponder on Portal The Portal fork of Ponder picked up upstream patches [0.17.6](https://github.com/subsquid-labs/portal-ponder/releases/tag/v0.17.6) and [0.17.8](https://github.com/subsquid-labs/portal-ponder/releases/tag/v0.17.8). Both are drop-in parity releases with no change to fork behavior. `npm install @subsquid/ponder` installs the newest; earlier versions stay installable by exact version. ## sqd-go [`sqd-go`](/en/sdk/alternative-clients/sqd-go) joins the alternative clients: a compiled Go indexer that streams EVM data from Portal and writes decoded events directly into ClickHouse. [mev.tools](https://github.com/mev-tools) built it for their own MEV and Polymarket indexing, workloads that need fast backfills and derived state such as balances and positions. * **Config-driven.** Contracts and events go in a `config.yaml`; `sqd-go` generates the tables, the decoders, and typed Go code. Checkpoint-based resume and reorg recovery are built in. * **Derived state in Go.** A custom processor compiled in with `--state` maintains state tables such as balances or positions from the same block stream. * **Backfill throughput.** In the maintainers' [open-indexer-benchmark](https://github.com/mev-tools/open-indexer-benchmark), the ERC-20 transfer events workload runs at about 2,900 blocks per second on default settings and about 7,500 with parallel fetching, with peak memory growing from 621 MB to about 2 GB in the tuned run. * **Community-maintained.** Not maintained by the SQD core team; bugs and feedback go to the [sqd-go issue tracker](https://github.com/subsquid-labs/sqd-go/issues). ## Dataset retirements completed The [67 datasets announced on August 6](#august-2026-dataset-retirements) were retired on schedule on August 20 at 12:00 UTC. ## Docs * **Portal error handling.** The Portal API [error contract](/en/portal/introduction/error-handling) is documented and surfaced in the API playground. * **Pipes error reference.** All 79 Pipes error codes now have pages in the [error reference](/en/sdk/pipes-sdk/reference/errors). * **Machine-readable API specs.** All seven per-chain Portal OpenAPI specs now carry unique operationIds, so function-calling adapters and agent tooling can target every operation. * **Cloud guide.** [Moving to a smaller database disk](/en/cloud/resources/shrink-database-disk) covers moving a squid's Postgres database to a deployment with a smaller disk, without downtime. ## August 2026 dataset retirements On **August 20, 2026 at 12:00 UTC**, we're retiring the 67 datasets listed below. They've seen very little usage over the past several months. We're consolidating what we maintain as we move from v2 gateways to Portal. * `0g-testnet` * `acala` * `aleph-zero` * `aleph-zero-evm-mainnet` * `arthera-mainnet` * `avail` * `basilisk` * `berachain-bartio` * `bifrost-kusama` * `bifrost-polkadot` * `bitfinity-mainnet` * `bitgert-mainnet` * `bridge-hub-kusama` * `bridge-hub-polkadot` * `canto` * `centrifuge` * `cere` * `clover` * `collectives-polkadot` * `crust` * `cyber-mainnet` * `darwinia` * `darwinia-crab` * `degen-chain` * `dfk-chain` * `dogechain-mainnet` * `equilibrium` * `frequency` * `gelato-arbitrum-blueberry` * `immutable-zkevm-mainnet` * `integritee` * `interlay` * `joystream` * `karura` * `khala` * `kilt` * `kintsugi` * `litentry` * `merlin-mainnet` * `moonsama` * `nakachain` * `neon-devnet` * `neon-mainnet` * `ozean-testnet` * `paseo` * `pendulum` * `people-chain` * `phala` * `picasso` * `plume-devnet` * `polkadex` * `polymesh` * `poseidon-testnet` * `reef` * `reef-testnet` * `robonomics` * `shibuya-substrate` * `shiden-mainnet` * `shiden-substrate` * `skale-nebula` * `sora-mainnet` * `subsocial-parachain` * `superseed-mainnet` * `tanssi` * `ternoa` * `turing-avail` * `turing-mainnet` What happens on the day, and what to do if you use one of these. ## Network availability updated As part of our ongoing transition from v2 to Portal, we retired 30 datasets with no significant measured demand and migrated 18 others to Portal-only access. The [network catalog](/en/data/all-networks) now reflects these changes. **Retired from Portal and v2** * `acurast-canary` * `aleph-zero-testnet` * `asset-hub-rococo` * `bitfinity-testnet` * `bitgert-testnet` * `bridge-hub-rococo` * `bridge-hub-westend` * `camp-network-testnet-v2` * `canto-testnet` * `collectives-westend` * `cyberconnect-l2-testnet` * `dancebox` * `dancelight` * `data-avail` * `dogechain-testnet` * `ethereum-holesky` * `foucoco` * `gemini-3h` * `immutable-zkevm-testnet` * `kyoto-testnet` * `merlin-testnet` * `neox-testnet` * `phala-testnet` * `polimec` * `rococo` * `rolimec` * `shibuya-testnet` * `stratovm-sepolia` * `superseed-sepolia` * `westend` **Available on Portal only** * `astar-mainnet` * `berachain-mainnet` * `binance-testnet` * `celo-mainnet` * `ethereum-hoodi` * `etherlink-mainnet` * `etherlink-shadownet` * `hyperliquid-mainnet` * `linea-mainnet` * `monad-testnet` * `moonbase-testnet` * `scroll-mainnet` * `scroll-sepolia` * `soneium-minato-testnet` * `taiko-mainnet` * `unichain-mainnet` * `zksync-mainnet` * `zora-mainnet` ## Pipes SDK 1.0 beta [`@subsquid/pipes` 1.0.0-beta.1](/announcements/pipes-sdk-1-0) is available on npm under the `beta` tag. * **Sources and targets.** Typed Portal streams cover EVM, Solana, Bitcoin, Tron, and Hyperliquid fills. Built-in targets write to ClickHouse, Postgres, BigQuery, and Parquet with cursor and reorg handling. * **CLI.** `@subsquid/pipes-cli` 1.0.0-beta.2 generates EVM and SVM projects and keeps the pipe ID stable when a project is regenerated. * **ClickHouse reorg recovery.** [`0.1.0-beta.18`](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v0.1.0-beta.18) lets a ClickHouse target restart after a reorg when the latest saved state has no finalized cursor. * **Postgres fork recovery.** [`1.0.0-alpha.17`](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-alpha.17) restores pre-fork row values, keeps finalized writes out of rollback batches, and handles block 0 as a valid finalized head. * **Multi-range queries.** [`1.0.0-alpha.20`](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-alpha.20) prevents false fork detection when a pipe resumes a query with disjoint ranges. It also accepts blocks without post-London or post-Cancun fields and exposes EIP-4844 blob gas fields on transactions. * **Tron field ranges.** [`1.0.0-beta.1`](https://github.com/subsquid/pipes-sdk/releases/tag/pipes-v1.0.0-beta.1) no longer rejects valid signed amount or timestamp values returned by Portal. * **Pipes SDK docs for three more data types.** New quickstarts and references cover [Bitcoin](/en/sdk/pipes-sdk/bitcoin/quickstart), [Tron](/en/sdk/pipes-sdk/tron/quickstart), and [Hyperliquid fills](/en/sdk/pipes-sdk/hyperliquid/quickstart), including data setup, query building, and decoding. * **Multiple indexers in Pipes UI.** [Pipes UI](/en/sdk/pipes-sdk/evm/guides/basic-development/pipes-ui#what-it-shows) can now track multiple indexers in one dashboard, with progress, speed, memory use, and query details for each indexer. ## Ponder on Portal [`@subsquid/ponder` 0.17.4-sqd.1](https://github.com/subsquid-labs/portal-ponder/releases/tag/v0.17.4) has been released. * **Transaction data.** [`0.16.8-sqd.1`](https://github.com/subsquid-labs/portal-ponder/releases/tag/v0.16.8) stores missing access lists as `NULL`, limits access lists to EIP transaction types 1 through 4, and reports a clear error if Portal keeps returning HTTP `429`. * **Realtime resume.** [`0.17.0-sqd.1`](https://github.com/subsquid-labs/portal-ponder/releases/tag/v0.17.0) prevents duplicate factory child records when realtime sync resumes after finalizing the same block. * **Shared factories.** [`0.17.2-sqd.1`](https://github.com/subsquid-labs/portal-ponder/releases/tag/v0.17.2) preserves child addresses discovered in realtime when a factory is shared across handlers. Rows returned by database operations are no longer changed by later writes to the same row. * **Realtime memory.** [`0.17.3-sqd.1`](https://github.com/subsquid-labs/portal-ponder/releases/tag/v0.17.3) fixes retention of unfiltered transactions during realtime sync. * **Large factory indexes.** Historical log and account transaction queries are split into smaller requests when their address filters exceed Portal's request limit. * **Portal realtime.** Experimental Portal realtime keeps oversized log filters on `/stream` by querying matching topics and filtering the results locally. * **Supported package lines.** The large factory fixes are also available in `0.15.17-sqd.3`, `0.16.6-sqd.3`, and `0.16.10-sqd.2`. ## Squid SDK The [July 30 release](https://github.com/subsquid/squid-sdk/releases/tag/2026-07-30) adds the `@subsquid/squid-sdk` package for using SDK data sources and utilities through one package. * **Portal and RPC fallback.** EVM processors can use Portal, RPC, or an automatic fallback across both. Source metrics are registered by the processor without separate setup. * **Database schemas.** `DB_SCHEMA` can set the PostgreSQL schema used by configuration, migrations, TypeORM, and GraphQL. * **Data source fixes.** Bounded EVM Portal streams now stop at the requested `to` block. The Hyperliquid fills service reopens a subscription when it keeps returning stale blocks. * **Squid SDK guides.** The docs now include quickstarts, how-it-works guides, and FAQs for EVM, Solana, Substrate, Fuel, Tron, and Starknet. Examples use current package versions. ## Portal v0.11 and v0.12 * **Overload handling.** [Portal v0.11.0](https://github.com/subsquid/sqd-portal/releases/tag/v0.11.0) rejects excess work before it slows active queries and uses HTTP `529` for overload instead of `503`. * **Stream correctness.** [Portal v0.11.1](https://github.com/subsquid/sqd-portal/releases/tag/v0.11.1) prevents a chunk from being returned twice when its workers are briefly unavailable. * **Bounded streams and reconnects.** [Portal v0.11.2](https://github.com/subsquid/sqd-portal/releases/tag/v0.11.2) caps buffered response parts per chunk. [Portal v0.11.3](https://github.com/subsquid/sqd-portal/releases/tag/v0.11.3) waits before reconnecting to a recovering worker. * **Query diagnostics.** [Portal v0.11.4](https://github.com/subsquid/sqd-portal/releases/tag/v0.11.4) identifies the worker behind each failed query attempt. * **Trace queries and self-hosting.** [Portal v0.11.7](https://github.com/subsquid/sqd-portal/releases/tag/v0.11.7) supports the current EVM trace API and adds self-hosting documentation. * **Upstream timeouts.** [Portal v0.11.8](https://github.com/subsquid/sqd-portal/releases/tag/v0.11.8) times out stalled Hotblocks, dataset configuration, and worker storage requests. Non-ASCII request IDs no longer close the connection. * **Real-time robustness.** [Portal v0.11.9](https://github.com/subsquid/sqd-portal/releases/tag/v0.11.9) retries a real-time request if its replica connection closes before any response bytes arrive, reports the correct head when dataset variants finalize independently, and falls back to the archival head when a real-time head is unavailable. * **Structured errors.** [Portal v0.12.0](https://github.com/subsquid/sqd-portal/releases/tag/v0.12.0) returns errors in a consistent envelope with separate `type` and `code` fields. Clients that parse plain text errors or a top-level `message` field must now read `error.message`. Overload responses include a retry hint. * **Capacity metrics and error handling.** [Portal v0.12.1](https://github.com/subsquid/sqd-portal/releases/tag/v0.12.1) counts refused streams by capacity constraint and exposes stream occupancy, saturation time, and the configured stream limit for alerts. The served API reference now explains the v0.12 error taxonomy, retry behavior, `Retry-After`, and request IDs. * **Portal capacity.** Portal deployments can now run multiple instances under one network identity, allowing more capacity without interrupting worker connections. ## Documentation * **Docs design.** The docs now use the sqd.dev design system across navigation, code blocks, tables, cards, and interactive queries. ## Network worker 2.14.0 * **Worker release**. [Worker 2.14.0](https://github.com/subsquid/worker-rs/tree/v2.14.0) hardens the transport layer, handles requests across multiple threads, and adds an experimental query engine. Workers before 2.13.0 are no longer supported. * **Stale RPC detection**. The Network now marks a worker as `stale_rpc` when it stops refreshing onchain configuration or reads it from a stale RPC. The Network App shows the worker as jailed, while worker logs and metrics report it as unreliable. ## Ponder on Portal, now documented The [Ponder on Portal](/en/sdk/alternative-clients/ponder) docs are live. Back your Ponder indexer's historical sync with SQD Portal while keeping your handlers, schema, and realtime behavior unchanged. * **[Quickstart](/en/sdk/alternative-clients/ponder/quickstart).** Run a Ponder indexer backed by Portal. * **[Migrate an app](/en/sdk/alternative-clients/ponder/guides/migrate).** Move an existing Ponder app onto the Portal-backed package and validate it. * **[Benchmarks and validation](/en/sdk/alternative-clients/ponder/benchmarks-and-validation).** Published results, test conditions, and correctness evidence. For the release details, see [Ponder on Portal, plus Pipes SDK updates](/changelog#ponder-on-portal-plus-pipes-sdk-updates). ## SQD is now in the Claude Directory The SQD connector is [live in the Claude Directory](https://claude.ai/directory/connectors/sqd). Add it to Claude in one click to query blockchain data across 200+ networks in natural language, with no custom connector setup and no API keys. * **One-click connect**. Open the [SQD listing](https://claude.ai/directory/connectors/sqd) and select **Connect**. It works across Claude on web, desktop, and mobile. See the [Claude Connector guide](/en/ai/claude-connector) for the full walkthrough. * **Powered by the Portal MCP server**. The connector runs the [Portal MCP server](/en/ai/mcp-server) (`https://portal.sqd.dev/mcp`), so Claude gets the full tool set for EVM, Solana, Bitcoin, Substrate, and Hyperliquid: discovery, transactions, logs, token transfers, and analytics. * **Ask in plain language**. "Latest USDC transfers on Base", "summarize this wallet on Ethereum", or "compare daily transactions on Arbitrum and Optimism". Claude picks the right tool, runs the query, and returns results with suggested follow-ups. ## Richer EVM data on Ethereum and Polygon We reindexed `ethereum-mainnet` and `polygon-mainnet` from genesis to expose more EVM fields on Portal. * **New block columns**. `uncles`, `withdrawals`, `withdrawalsRoot`, `parentBeaconBlockRoot`, and `requestsHash`. * **New transaction columns**. `accessList`, `logsBloom`, `blobGasUsed`, and `blobGasPrice`. * **Queryable across full history**. Both datasets were re-synced from scratch, so the complete column set is available from genesis to the head. *** ## Robinhood Chain on Portal [Robinhood Chain](/en/data/evm/robinhood-mainnet) is available on SQD Portal as a private dataset. It's an Arbitrum Orbit (Nitro) EVM L2 with real-time data. * **Network**. `robinhood-mainnet`, chain ID `4663`. * **Private access**. Data access requires a custom arrangement. [Talk to our team](https://calendly.com/t-tyrie-subsquid/30min) about dataset scope, pricing, and onboarding. ## Network retirements * **Botanix retired.** Affected networks: `botanix-mainnet`, `botanix-testnet`. * **MemeCore retired.** Affected networks: `memecore-mainnet`, `memecore-insectarium`, `formicarium-testnet`. * **Exosama retired.** Affected network: `exosama`. ## Ponder on Portal, plus Pipes SDK updates ### `@subsquid/ponder` A drop-in [Ponder](https://ponder.sh) build that runs historical backfill over SQD Portal. Same handlers, schema, and `ponder` binary; add one `portal:` line per chain. Published as `0.16.7-sqd.1` (latest), `0.16.6-sqd.2`, and `0.15.17-sqd.2`. * **Faster backfill**. Full single-chain history indexed about 3.6x faster on the same host. A 15-chain Euler config indexed 28.4M events in 51m 47s (\~9.1k events/s) with zero public RPC calls. * **Correctness fixes**. Factory-app children now persist across restarts, and silent gaps from lagging replicas ("phantom coverage") are eliminated. * **Migration**. Run `npm i @subsquid/ponder`, then add `portal: "https://portal.sqd.dev/datasets/"` per chain. Realtime stays on your own RPC (experimental `PORTAL_REALTIME=stream`). ### Pipes SDK * **Tron support**. Build Pipes SDK indexers on `tron-mainnet`. * **Parquet target**. Write pipeline output directly to Parquet files. ## Tron on Portal Tron Mainnet is now available on SQD Portal as a real-time dataset, queryable from genesis. The query model covers blocks, transactions, TVM event logs, and internal transactions, with dedicated filters for native TRX transfers, TRC-10 asset transfers, and `TriggerSmartContract` calls. * **Endpoint**. Stream from `https://portal.sqd.dev/datasets/tron-mainnet`. The [Tron data page](/en/data/tron/tron-mainnet) has the schema and a quick-start query. * **TRC-20 and smart contracts**. Match TRC-20 `Transfer` events through `logs` (the TVM is EVM-compatible, so they share the ERC-20 topic signatures), or filter `triggerSmartContractTransactions` by contract address and function selector. * **Schema and playground**. Browse the [OpenAPI schema](/en/ai/tron-openapi) and run requests in the [API playground](/en/api/tron/stream). ## Network worker 2.13.0 [Worker 2.13.0](https://github.com/subsquid/worker-rs/tree/v2.13.0) supports new EVM trace filters and reports the current epoch so the Network can detect workers using stale RPC data. ## Portal MCP Server v0.7.9 The [Portal MCP Server](/en/ai/mcp-server) focuses on developer and agent ergonomics for natural blockchain questions. * **Entity resolution**. New `portal_resolve_entity` tool turns token symbols, contract aliases, pool IDs, protocol names, and Hyperliquid coin names into query-ready addresses and filters. Agents resolve "USDC on Base" or "Uniswap v3" to concrete filters before building a query, instead of hardcoding addresses. * **Token-list lookups**. Symbol-based queries resolve through open-source token lists from CoinGecko, DefiLlama, and similar public sources rather than a fixed set of built-in constants, so more tokens work without manual address entry. * **Natural-language time windows**. Query tools accept phrasing like "past 30 minutes" and "last hour" across more parameters, on top of the existing `timeframe` values. * **Investigation guidance in responses**. Results now include suggested investigation paths with primary evidence fields, pivot fields, and follow-up filters, so agents can trace an incident across tools without guessing the next call. * **Cross-VM parity**. Added regression coverage for Solana, Bitcoin, and Hyperliquid investigations, alongside a unified response envelope, wallet fund-flow triage, and hardened ERC-721 mint lookups. ## Richer, more accurate EVM data Based on customer requests, we expanded EVM coverage with new post-Shanghai and post-Cancun fields, and tightened the correctness and completeness of our datasets. * **Validator withdrawals (post-Shanghai)**. Blocks now expose `withdrawals`, with `index`, `validatorIndex`, `address`, and `amount` per entry, plus `withdrawalsRoot`. Live on Ethereum mainnet and Gnosis, with BSC rolling out next. * **Blob gas accounting (EIP-4844)**. On Ethereum mainnet, blocks expose `blobGasUsed` and `excessBlobGas`, and type-3 blob transactions expose `blobGasUsed`, `blobGasPrice`, `maxFeePerBlobGas`, and `blobVersionedHashes`. * **Corrected `selfdestruct` traces**. We reindexed `ethereum-mainnet` so `selfdestruct` traces no longer carry an empty `refundAddress`. * **More complete Polygon data**. We backfilled the missing Bor state-sync transactions and corrected the transaction counts and log indexes on the affected blocks. * **Cleaner decoded events**. We standardized how empty `args` and `named_args` are represented so queries behave consistently, and confirmed that no decoded events are dropped when a real-time indexer restarts. Behind this work, we continuously cross-check our datasets against independent sources, both node RPC and external reference datasets, comparing row counts and field-level values to catch discrepancies early. *** ## Query Portal over JSON-RPC (beta) An RPC-compatible endpoint over SQD Portal is in early beta, rolling out to select customers. It lets an EVM indexer talk to Portal over standard JSON-RPC, answering the data-heavy historical methods (`eth_getLogs`, blocks, `eth_getBlockReceipts`, and traces) straight from Portal, so it works as a drop-in RPC for an indexer's sync and backfill path without any code changes. * **Portal-backed read methods**. A fast Rust JSON-RPC service serves the indexer sync path (logs, blocks, receipts, and traces) from Portal data instead of from node providers. * **Faster, with zero errors and fallback built in**. In a 1,000-block Ethereum backfill, the Portal-backed path ran about 23% faster than a standard node (76s vs 99s), while returning zero errors and having an independent fallback provider already wired in. * **No single-provider lock-in**. Methods Portal doesn't serve directly (state calls like `eth_call` and `eth_getBalance`, realtime subscriptions, transaction submission) route transparently to that independent provider, so you get a complete RPC plus redundancy. Building something that needs this? [Get in touch with us](https://calendly.com/t-tyrie-subsquid/30min). ## Network worker 2.12.0 [Worker 2.12.0](https://github.com/subsquid/worker-rs/releases/tag/v2.12.0) fixes query simplification and updates the network library to support larger responses. ## Portal v0.9.12, v0.10.0, and v0.10.1 * **Query validation.** [Portal v0.9.12](https://github.com/subsquid/sqd-portal/releases/tag/v0.9.12) returns `400 Bad Request` when `buffer_size` or `max_chunks` is zero and when a response exceeds the worker response limit. * **Larger responses.** [Portal v0.10.0](https://github.com/subsquid/sqd-portal/releases/tag/v0.10.0) accepts larger responses from Network workers. * **Graceful restarts.** [Portal v0.10.1](https://github.com/subsquid/sqd-portal/releases/tag/v0.10.1) stops accepting new requests, completes active requests, and closes cleanly during a restart. ## Retirements and dataset changes As we consolidate onto Portal (see the [gateway API keys announcement](/announcements/gateway-api-keys)), we're retiring several legacy gateways and datasets. * **Goerli testnet gateways retired (May 21)**: Ethereum, Arbitrum, Optimism, Base, Mantle, Polygon zkEVM, Zora Goerli, and Polygon Mumbai. * **Gateways and datasets removed**: Fuel, Starknet, Eclipse, Fantom, and CrossFi. * **Solana Mainnet v2 retention reduced**. The v2 archive now keeps only the most recent 30 days. ## API keys for gateways (v2 Archives) External requests to gateway endpoints (v2 Archives) now require an API key. Sign in at [portal.sqd.dev](https://portal.sqd.dev) to generate one. SQD Cloud users are unaffected. * **Generate keys at [portal.sqd.dev](https://portal.sqd.dev)**. Sign in with GitHub, Google, or email, and a `default` key is minted on first sign-in, shown with a copy button and a ready-to-paste `curl` example. You can also create additional named keys with an optional TTL and revoke them at any time. * **Per-key usage**. Track the requests and bytes served per key over the last 7 days. * **SDK support**. The Squid SDK sends your key automatically (`SQD_API_KEY`), surfaces friendly errors, and emits a sunset notice (`X-SQD-Sunset`) for unauthenticated access. See [Accessing SQD data with API keys](/en/data/api-keys). * **Enforcement was on (May 19, 12:00 UTC)**. Unauthenticated gateway requests now return `401`. Keys apply to gateway endpoints only; they aren't required for Portal yet. Squids using RPC (no `.setGateway()`) are unaffected. We've also rewritten the [gateway-to-Portal migration guide](/en/sdk/migration/gateway-to-portal): EVM and Solana now live on one page, with skills that let AI agents automate the migration. ## Network worker 2.11.1 [Worker 2.11.1](https://github.com/subsquid/worker-rs/releases/tag/v2.11.1) fixes a StateManager deadlock, separates chunk identity from chunk names, and turns query processor panics into recoverable errors instead of stopping the worker. ## Network worker 2.11.0 [Worker 2.11.0](https://github.com/subsquid/worker-rs/releases/tag/v2.11.0) supports the new onchain contracts and EVM dataset columns. Fractional compute-unit calculations give lightweight queries more precise capacity accounting. ## Typed stream APIs in Squid SDK The [May 11 Squid SDK release](https://github.com/subsquid/squid-sdk/releases/tag/2026-05-11) adds typed EVM and Solana stream sources and request builders. * **EVM codecs and typegen.** New major versions of `@subsquid/evm-abi`, `@subsquid/evm-codec`, and `@subsquid/evm-typegen` use the typed codec API and generate importable ABI modules with typed event topic filters. * **Unfinalized streams.** `@subsquid/batch-processor` accepts unfinalized data sources. * **Recovery and retries.** The TypeORM store improves hot block rollback and recovery. HTTP `408`, connection resets, and lost RPC subscriptions are retried. ## Portal MCP Server v0.7.8 The [Portal MCP Server](/en/ai/mcp-server) ships a focused release on query correctness, faster analytics, and release discipline. * **EVM contract deployment lookup**. New `portal_evm_get_contract_deployment` tool returns the deployment block, transaction, and deployer for any EVM contract address. * **EVM investigator queries**. `portal_evm_query_transactions` now supports `transaction_type` and `scan_order`, plus method/event aliases, top-N rankings by value or gas, and richer client-side filters. Prompts like "first Ethereum mainnet tx of type 0x1 from block 12,244,000" work in one call. * **Faster analytics**. Short-lived query caching and larger safe EVM time-series chunks cut redundant Portal scans on repeated calls. * **Real-time timestamp windows**. Relative timeframes now anchor to the actual latest indexed block or slot timestamp on real-time datasets, including Solana and Hyperliquid fills. No more guessing which block corresponds to "now". Both real-time and historical windows resolve to exact timestamps. Prompts like: * "What's been happening in the past 48 minutes on Base?" * "Show Solana activity from the last 10 minutes." * "List Hyperliquid fills between 2026-04-15 09:00 UTC and 09:30 UTC." resolve to precise block or slot ranges instead of best-effort estimates. ## AI Agent Skills v1.1.4 [Agent Skills](/en/ai/agent-skills) updated to track the v0.7.8 MCP tool surface. * **Portal skill retargeted** to Portal MCP v0.7.8 tool names. Earlier flat names like `portal_query_logs` are now `portal_evm_query_logs`. The skill flags compatibility so older clients fall back to raw API calls. * **Timestamp-to-block guidance for real-time data**. The Portal skill now documents the `/timestamps/{unix-seconds}/block` endpoint and `portal_debug_resolve_time_to_block` MCP tool for resolving any timestamp (including hotblocks-era timestamps from minutes ago) to a precise block or slot. Agents handling prompts like "transfers on Base in the last 4 hours" or "Solana activity from the last 10 minutes" stop guessing block ranges and ask the dataset directly. Works on all Portal datasets. * **New `squid-perf` skill**. Profile and compare sync-time performance across one or more Squid SDK deployments. Pulls logs via the `sqd` CLI, generates a self-contained HTML report and a Markdown summary with wall-clock, active-time, and downtime breakdowns at log-spaced block breakpoints. ## Adding API keys to gateways The first change to gateways since we opened up public datasets in 2023. From **Tuesday, May 19 at 12:00 UTC**, external calls to gateway endpoints will need an API key. If you're on SQD Cloud, this won't affect you. If you're self-hosting a squid or otherwise calling gateways from outside of SQD Cloud, [generate an API key](/en/data/api-keys) and attach it to your requests. That's the only change. Looking ahead: [Portal](/en/portal/overview) is our priority, and gateways will be sunset later this year, with some datasets being deprecated entirely before then. API keys are coming to Portal next, alongside pricing tiers. Timing, who's affected, and how to get a key. *** ## Docs 2.0 is live After a few months in beta, the new docs are out of beta. [docs.sqd.dev](https://docs.sqd.dev) now points here. A few highlights from the past few months: ### New content * **Full [Pipes SDK](/en/sdk/pipes-sdk/evm/quickstart) guide series.** Pipe anatomy, cursor management, fork handling, stateful transforms, EVM contract event processing, and factory patterns. Both EVM and Solana. * **Pipes CLI.** `pnpx @subsquid/pipes-cli@1.0.0-alpha.4 init` scaffolds a runnable pipe in a few prompts. Walkthrough in the [quickstart](/en/sdk/pipes-sdk/evm/quickstart). * **[Solana gateway to Portal migration guide](/en/sdk/migration/gateway-to-portal#solana)** with a height-to-slot converter. * **[Bitcoin on Portal](/en/portal/bitcoin/overview)**: queries, datasets, OpenAPI schema. * **Portal stream semantics** documented end-to-end: continuation, reorgs, finality. ### AI-ready by default * **[Documentation MCP server](/en/ai/mcp-server-docs)** for searching docs from Claude or Cursor without leaving your editor. * **[Portal MCP server](/en/ai/mcp-server)** for querying blockchain data in plain English. * **OpenAPI schemas** for EVM, Solana, Substrate, Bitcoin, Hyperliquid, and HyperCore. * `llms.txt`, `llms-full.txt`, and `skill.md` published for AI tooling integration. ### Easier to find things * Build menu now organized by ecosystem (EVM, Solana, Substrate, Other) instead of by SDK. * Portal API navigation split into **Reference** and **Working with** so you can scan instead of dig. * [AI section](/en/ai/ai-development) consolidated. * Network tables with clearer columns and direct links to chain pages. ### Polish * Introduced dark mode. * Every page has a proper Open Graph preview when shared. * Hundreds of broken links fixed and missing descriptions added across the site. * Per-chain network pages are now generated from a Portal datasets API sync. Adding a chain to Portal will update the docs automatically in the future. ## Portal * **Timestamps resolve in hotblocks too.** Timestamp-to-block resolution now works against live hotblocks data, not just archival. Timestamp-based queries work end-to-end. * **Cleaner archival to hotblocks handoff.** Queries that cross the boundary between archival data and live hotblocks now stitch results together instead of failing at the gap. * **HotblocksDB scales independently** of the Portal frontend, removing a bottleneck on high-throughput chains. * **Transport layer hardening.** Flaky peers and bootnode disconnects no longer drop active streams. ## Status page A first version of a public status page is in place. It's an internal foundation for now, not aimed at external users yet. Per-dataset uptime, ingestion lag, and Portal availability are tracked centrally. ## Stronger data consistency on Ethermint-based chains Older Ethermint versions suffer from a cluster of data inconsistencies that every downstream consumer inherits from the node, including the same transaction hash appearing in multiple blocks. That issue is acknowledged in the [Cronos v1.7.1 release notes](https://github.com/crypto-org-chain/cronos/releases/tag/v1.7.1) and explicitly left to indexers to reconcile. We now detect and resolve these issues at ingestion time. The first beneficiary is our Cronos historical dataset, but the approach generalizes to every Ethermint-based network. Downstream consumers can now rely on fundamental invariants like unique tx hashes, receipts aligned with block headers, and verifiable logs that the raw RPC output alone cannot guarantee. ## Datasets * **Arbitrum**: transaction type `0x65` ingests correctly now. * **Solana historical**: missing instructions backfilled. * **Portal API metadata fix**: 45 mainnets that were mislabeled as testnets now report the right network type. * Sync issues fixed on `gnosis-mainnet` archival. ## Network economics * **Auto-buyback of SQD tokens** wired into Portal Pools smart contracts. * Distribution rate calculation in Portal Pools contracts corrected. ## SDK * **`@subsquid/typeorm-migration`**: `down()` migrations now reverse the order of operations, so drops happen in a safe order. * **`@subsquid/substrate`**: Westend Asset Hub runtime `westmint@1022000` no longer crashes the call parser. ## Network App & Portal Pools UI * **Portal Pool staking** surfaced on [network.sqd.dev/assets](https://network.sqd.dev/assets). * **Withdraw-from-failed-pool** now supported. * Color and state alignment plus several UI fixes from Telegram-reported issues. ## sqd.dev Website Launch The new [sqd.dev](https://sqd.dev) is live. A complete rebuild from scratch, migrated off Webflow. Faster, cleaner, and fully owned in-house. ### Portal Enterprise Service Agreements Formal Enterprise Service Agreements are now available, with defined SLAs, uptime commitments, data freshness guarantees, tiered chain support, and dedicated incident management. [Get in touch](https://sqd.dev/contact/) to learn more. ### Portal Pools. Lambda Revenue Pool The first public revenue pool, **Lambda x SQD**, launched with a 5M SQD capacity and filled within hours of going live. ### Cloud. PostgreSQL version support Squids now support configurable PostgreSQL versions 14, 15, 16, 17, and 18, with the default upgraded to **PostgreSQL 18**. Set your preferred version in `squid.yaml` under the `postgres` addon. Two new session control inputs were also added: `idle_in_transaction_session_timeout` and `idle_session_timeout`. ### Portal MCP Server v0.7.5 → v0.7.7 The [Portal MCP Server](/en/ai/mcp-server) got a significant overhaul across three releases: * **Tool catalog overhaul**. Audited every public tool and consolidated down to \~20 core tools (23 with Substrate), with a unified description style, input schema, response envelope, and routing metadata. * **Substrate support**. Three new tools (`portal_substrate_get_analytics`, `portal_substrate_query_events`, `portal_substrate_query_calls`) bring Polkadot- and Kusama-style chains to parity with EVM, Solana, Bitcoin, and Hyperliquid. * **Interactive panel UI (v1)**. New `_ui` envelope (panels, metric cards, follow-up actions) renders results as a dashboard in MCP-aware clients instead of raw JSON. * **LLM next-step hints**. Every response now includes structured follow-up guidance so multi-step investigations work without hand-holding. * **Experimental EVM OHLC tool**. `portal_evm_get_ohlc` builds candles from Uniswap v2/v3/v4 and Aerodrome Slipstream swaps with a recent-trades tape. * **BTC and Hyperliquid tools**. Available since v0.7.5. ### Substrate OpenAPI Schema Published a public OpenAPI schema for Substrate datasets, unlocking better developer tooling and documentation for Polkadot ecosystem data. ### New Chains & Datasets * **MegaETH**. Hotblocks support added for both mainnet and testnet. * **ADI**. Added as a hotblocks-only dataset. * **Alpen**. Onboarded (chain ID 8150) with archival data. * **Hyperliquid `replica_cmds`**. Both S3 archival and real-time ingestion are now live. * **Solana token balance siblings**. New `transactionTokenBalances` and `transactionBalances` flags for the `tokenBalances` filter, letting users retrieve all token balance updates for the parent transactions of matching balances without overfetching. * **Aptos & Aleo**. Initial onboarding research completed. ### Discontinued Substrate Chains Following a review of ecosystem usage, the following Substrate datasets have been discontinued and removed from the docs and network: **Amplitude**, **InvArch** (parachain and Tinkernet), and **Zeitgeist** (mainnet and testnet). ### Bug Fixes * **Better errors for missing columns**. "Column not found" errors from the query engine now correctly return HTTP 400 (Bad Request) instead of a generic 500, making queries much easier to debug. * **Hotblocks memory optimization**. Improved memory management in hotblocks services with usage limits and proper cleanup for stable performance under high load. ### Squid SDK * **ESM relation types.** The [April 15 Squid SDK release](https://github.com/subsquid/squid-sdk/releases/tag/2026-04-15) generates TypeORM relation fields with `Relation<>` and re-exports the type for ESM projects. * **RPC retries.** EVM processors retry temporarily non-canonical block responses and out-of-range log errors. ## Squid SDK * **Transaction isolation.** The [April 8 Squid SDK release](https://github.com/subsquid/squid-sdk/releases/tag/2026-04-08) lets GraphQL and OpenReader APIs use `SERIALIZABLE`, `REPEATABLE READ`, or `READ COMMITTED` database transactions. * **Finalized head handling.** The Portal client ignores a finalized head that moves backward during a stream. ## Network worker 2.10.1 [Worker 2.10.1](https://github.com/subsquid/worker-rs/releases/tag/v2.10.1) raises the incoming query limit from 256 KB to 4 MB. Worker versions 2.10.0 and earlier were subsequently retired. ## Bitcoin on Portal Bitcoin on SQD Portal Bitcoin is now fully available on the public portal. Archival data, real-time hotblocks, and all. This makes SQD the first decentralized data network to serve Bitcoin block and transaction data through the same Portal API used for EVM and Solana chains. Timeline: * **Feb 23**. Archival data synced, Bitcoin query support added to Portal * **March 19**. Versioned datasets introduced, majority of workers upgraded to >2.9 * **March 24**. Dataset and hotblocks deployed to public portal Explore the Bitcoin Portal API. Query blocks, transactions, inputs, and outputs with interactive examples ### Portal & Network * **Network worker 2.10.0**. [Worker 2.10.0](https://github.com/subsquid/worker-rs/releases/tag/v2.10.0) adds Bitcoin datasets, Tempo fields, advanced query execution statistics, and connectivity fixes. * **Versioned datasets**. New mechanism allowing datasets to be restricted to workers running a minimum client version, enabling safe rollout of new chain support * **Worker execution time breakdown**. Worker logs now show per-step timing (query parsing, execution, compression, signing, serialization) * **Worker auto-reconnect**. Workers now automatically reconnect if initial network bootstrapping fails * **Error propagation**. Query engine errors like `"column not found"` now return proper 400 Bad Request instead of 500 Server Error * **Parent hash check fix**. Fixed incorrect parent block hash validation in EVM datasets * **Head monitor log fix**. Fixed log spam (2.5M logs/30min) * **SQL query engine**. Integrated into the network for future SQL query support ### Portal MCP Server v0.7.5 The [Portal MCP Server](/en/ai/mcp-server) reached v0.7.5 with 29 tools covering EVM, Solana, Bitcoin, and Hyperliquid data. All queryable in natural language through Claude, Cursor, or any MCP-compatible AI assistant. Since the last update: * **Bitcoin support**. 5 new tools for querying Bitcoin transactions, inputs, outputs, analytics, and time series * **Tool consolidation**. Consolidated from 41 to 29 tools by converting 13 standalone tools (traces, state diffs, NFT transfers, gas analytics, etc.) into include flags on core query tools * **Per-chain analytics and time series**. New analytics and time series tools for Solana, Bitcoin, and Hyperliquid with chain-specific metrics (TPS, fee trends, volume, trader rankings) * **Retry logic**. Added 503/409/429 retry with exponential backoff for production reliability * **Timeframe estimation**. Improved short timeframe handling; skips timestamp API for windows under 2h and uses per-chain block time estimates * **Fuzzy dataset resolution**. `"base"` resolves to `base-mainnet`, `"eth"` to `ethereum-mainnet` * **Timeframe parameter**. All query tools support `timeframe` (e.g., `"24h"`, `"7d"`) instead of manual block ranges * **Bounded response pages**. Query tools return a bounded page plus a `next_cursor` to keep responses within AI context windows, with a soft notice on heavy ranges * **Chunked fetching**. Hyperliquid and Solana queries use chunked fetching to prevent memory issues * **Known limitation**. Large unfiltered queries on high-TPS chains (Solana, Hyperliquid) can still hit memory limits; use tight filters and small block ranges on dense chains ### AI Agent Skills v1.1.1 * Successfully validated via 40+ indexers created by AI ([source](https://github.com/karelxfi/ai-pipes)) * Further reduced from 4 skills to 3. 2 for Pipes SDK, 1 for Portal * **Bitcoin support**. Added Bitcoin query structure, fields, filters, and dataset slug to the [AI skill file](/en/ai/skill-md) ### Network App * **Analytics as default tab**. Dashboard now opens to Analytics instead of Workers * **TVL chart upgrade**. Stacked chart showing workers, delegations, portal pools, and dedicated portals * **Dynamic chart ranges**. Charts now update to current date instead of ending at a fixed point ### Portal Pools UI * **Claims tab**. New Claims tab in the activity table alongside Withdrawals * **Activity separation**. Separated "all activities" from "my activity" views * **"All time" filter**. Added to charts ### Cloud * **Postgres session timeouts**. New manifest inputs `idle_in_transaction_session_timeout` and `idle_session_timeout` for controlling database session behavior * **Connection visibility**. Postgres connection count now exposed in Odyssey for monitoring and alerting * **Postgres addon warning**. Added warning that removing Postgres addon wipes the database * **Docker build optimization**. Reduced squid-sdk Docker image build times * **Pino logger support**. Squid SDK now supports Pino integration, enabling DataDog and other structured logging platforms ### SDK * **Custom processor metrics.** The [March 23 Squid SDK release](https://github.com/subsquid/squid-sdk/releases/tag/2026-03-23) lets `run()` use a custom Prometheus server and lets runner metrics use a custom block-count function. * **Deployment inspection.** [Squid CLI v3.3.1](https://github.com/subsquid/squid-cli/releases/tag/v3.3.1) adds `sqd view` for inspecting a deployment and opening its Cloud page. * **Pipes SDK 1.0.0 (Alpha 1)**. First alpha release of Pipes SDK 1.0.0 * **NPM trusted publishing**. Configured OIDC-based trusted publishing for all \~84 Squid SDK packages * **Pipes SDK `transaction.nonce`**. Changed to `bigint` to accept both string and int from Portal responses ### New Chains * **Bitcoin** added to Portal (archival + hotblocks) * **Tempo** hotblocks now live; added to private portals * **TAC** network onboarded ### Bug Fixes * **Bittensor testnet**. Fixed ingestion crash on EIP-7702 transactions * **Portal event data**. Fixed Arbitrum event data inconsistency * **Pings collector**. Fixed caching issues ## Portal MCP Server v0.6 The [Portal MCP Server](/en/ai/mcp-server) received a major update with new chain support and stability improvements: * **Hyperliquid support**. New tools for querying Hyperliquid data: `portal_query_hyperliquid_fills` (trade executions, PnL, fees) and `portal_query_hyperliquid_replica_cmds` (orders, cancels, transfers, leverage) * **Stability improvements**. Capped response sizes, lowered default limits, fixed stdout pollution breaking stdio transport, corrected field names, fixed HyperEVM alias resolution, added Cloudflare Workers compatibility ### AI Agent Skills v1.1 [AI Agent Skills](/en/ai/agent-skills) updated with better coverage and leaner skill definitions: * **Pipes SDK skill consolidation**. Reduced from 21 skills to 4 focused ones with reference docs * **Portal query skills**. New skills for EVM logs, traces, transactions, Solana instructions, and Hyperliquid data * **Factory pattern & proxy detection**. Added factory contract pattern documentation, proxy detection workflow, and timestamp handling ### Portal & Network * **Domain migration**. `app.subsquid.io` → `cloud.sqd.dev`, `network.subsquid.io` → `network.sqd.dev` * **Tempo query support**. Portal now supports Tempo network transaction queries * **503 error fix**. Resolved high-volume 503 errors on public portal * **Network stability**. Fixed bootnode connection drops, head monitor retry logic, and stable-mainnet chain stuck issue ### Docs * **Showcase**. New page highlighting live products built with SQD (since retired in favour of the [SQD ecosystem directory](https://sqd.dev/ecosystem)) ## Portal & Network Updates * **Tempo network fields**. Portal query layer supports Tempo network transaction fields * **Portal Pools dashboard**. Portal Pools UI integrated into Network App * **Portal streaming performance**. Fixed streaming bottleneck (was capped at \~5 MB/s) * **Portal MCP Server**. MCP server deployed to portal.sqd.dev ([learn more](/en/ai/mcp-server)) * **HyperCore documentation**. Added documentation for [Hyperliquid fills queries](/en/data/hyperliquid/hyperliquid-fills) and [replica commands](/en/data/hyperliquid/hyperliquid-replica-cmds) alongside [HyperEVM support](/en/portal/hyperliquid/overview) ### SDK & Improvements * **Tempo Transactions**. Added support for Tempo Transactions (type 0x76) in Squid SDK * **Squid SDK fix**. Fix duplicate charset in Prometheus content-type header ### Cloud Features * **Billing Limits & Usage Alerts**. Users can now set spending caps and receive usage alerts ([pricing details](/en/cloud/pricing/overview)) * **Cloud UI Refresh**. Refreshed UI for the [Cloud platform](/en/cloud/overview) ### New Chains * **World** added to Portal * **Sei** added to Portal * **Stable** added to Portal * **Katana** added to Portal * **Flare hotblocks** enabled * **Ink hotblocks** enabled ### Bug Fixes * **llms.txt and llms-full.txt**. Fixed issues with llms.txt and llms-full.txt endpoints; both now work correctly ### Removed * **Retention feature**. Retention functionality dropped from Portal API ## Portal & Data Features * **Dynamic metadata**. Portal now reads and exposes dataset metadata dynamically instead of hardcoded mapping * **`/resolve-timestamp` endpoint**. New Portal endpoint for converting Unix timestamps to block numbers ### SDK & Pipes * **Pipes UI metrics**. Metrics fetching moved to server-side, supports metrics display for multiple pipe instances * **Profiler**. Made OpenTelemetry compatible * **Production-ready Pipes SDK EVM example**. Production-ready pipe example for EVM published * **Relationships without FK constraints**. Squid SDK now supports relationships without foreign key constraints * **ABI service**. ABI service for simplified ABI fetching and type generation shipped * **Solana instruction decoder**. Added to Pipes SDK with rollback chain logic improvements ### Cloud Improvements * **Read-only database access**. Cloud deployments now support read-only database access; Viewer role gets read-only DB access by default * **Latency chart**. Latency chart added to [Cloud deployment metrics](/en/cloud/resources/monitoring); latency data exposed in API * **CLI confirmation prompts**. `sqd` CLI now prompts for confirmation before deleting PostgreSQL databases * **Advanced logs filtering**. Advanced [logs filtering](/en/cloud/resources/logging) and live log streaming in Cloud console * **RPC proxy enhancements**. ERPC third-party plugin added, request batching added to [RPC proxy](/en/cloud/resources/rpc-proxy) ### Network & Other * **Network worker 2.7.0**. [Worker 2.7.0](https://github.com/subsquid/worker-rs/releases/tag/v2.7.0) adds SQL query support. * **Network worker 2.8.0**. [Worker 2.8.0](https://github.com/subsquid/worker-rs/releases/tag/v2.8.0) adds Hyperliquid replica command queries. * **Network App sidebar**. Sidebar menu in Network App is now collapsible * **Portal fallback**. Fallback handling for when Portal is down * **Botanix & Gnosis hotblocks**. Botanix added to Portal, Hotblocks enabled for Gnosis Mainnet * **Botanix docs**. Botanix network added to documentation ### Community & Ecosystem * **Agent Skills on skills.sh**. SQD agent skills published to [skills.sh](https://skills.sh/?q=subsquid), making them discoverable in the AI agent skills marketplace * **Agentic Commerce Market Map**. SQD featured in [Agentic Commerce Market Map](https://x.com/helloSQD/status/2027079167880396966) as blockchain data infrastructure for AI commerce agents * **Polygains**. Community member launched [Polygains](https://polygains.com/), a MEV analytics platform built with SQD ([announcement](https://x.com/helloSQD/status/2023802079769489905), [source code](https://github.com/mev-tools/polygains/)) * **Solana Data Room**. Interactive Solana network statistics dashboard launched at [room.soldexer.dev/feb-2026](https://room.soldexer.dev/feb-2026). The network, in one glance ([announcement](https://x.com/helloSQD/status/2021625892913787388?s=20)) ## AI-ready docs and tooling If you use Claude, Cursor, or any AI coding assistant. You can now query blockchain data and search our docs without leaving your editor. * **Portal MCP Server**. Add it to your Claude or Cursor config and your AI assistant gets direct access to SQD Portal. Query 100+ chains, pull token transfers, check wallet activity, or analyze contracts in plain English. [Set it up →](/en/ai/mcp-server) * **Docs MCP Server**. A separate MCP server that lets your AI assistant search and retrieve SQD documentation directly. Answers grounded in our actual docs, not whatever the model was trained on. [Learn more](/en/ai/mcp-server-docs) * **llms.txt and SKILL.md**. Machine-readable files that help AI tools understand what SQD does. Compatible tools pick these up automatically. ([llms.txt](/en/ai/llms-txt), [SKILL.md](/en/ai/skill-md)) * **Context7**. Our docs are indexed on [context7.com](https://context7.com). If your AI tools integrate with Context7, you already have SQD docs in context. * **OpenAPI schemas**. [EVM](/en/ai/evm-openapi) and [Solana](/en/ai/solana-openapi) OpenAPI schemas exposed in documentation ## Portal & Network * **Real-time datasets**. Real-time data now available for: `arbitrum-sepolia`, `optimism-mainnet`, `monad-testnet`, `monad-mainnet`, `gnosis-mainnet` * **Real-time APY & TVL charts**. APY and TVL charts now update in real-time * **Dynamic dataset sizes**. Dataset sizes now dynamically displayed on SQD.ai, updating every \~24 hours * **Portal API V2 compatibility dropped**. V2 compatibility layer removed from Portal API specification * **`x-sqd-head-number` header**. Portal now supports `x-sqd-head-number` header for checking latest block ### SDK Features * **ABI service integration**. ABI service now integrated directly into the Typegen CLI * **Topic filtering**. Topic filtering support added for Factory contracts and evmDecoder * **EVM decoder enhancements**. Contract address filtering, ZSTD support, factory topic filtering in Pipes SDK * **Squid SDK improvements**. Use Etherscan proxy for typegen CLI, missing export for `InstructionRecord`, handle `query exceeds max results` error ### Network * **Portal Pools contracts**. Portal Pools smart contracts deployed to Arbitrum mainnet (PortalRegistry, PortalPoolFactory, FeeRouterModule) * **Etherlink migration**. Etherlink archives migrated from Ghostnet to Shadownet ### New Chains & Hotblocks * **megaETH Testnet** added to v2 archives and Portal * **MegaETH Mainnet** network page added to documentation * **Hotblocks enabled** for Monad, Optimism Mainnet, and Arbitrum Sepolia ## Portal Pools Whitelist Launch [Portal Pools](https://network.subsquid.io/portal-pool/0x438c2a47e82cd445524ce5651ae7e6c1dd386d09) went live with the first [whitelist-only release](https://network.subsquid.io/portal-pool/0x438c2a47e82cd445524ce5651ae7e6c1dd386d09). Smart contracts deployed on Arbitrum mainnet, whitelisted community members staking SQD in exchange for USDT rewards. The first pool was oversubscribed within 3 days. ### Portal & Network * **ZSTD \~30% faster**. Switching from gzip to zstd compression sped up Portal responses by \~30% in benchmarks * **Solana query relations**. `transactionTokenBalances` and `transactionBalances` flags added to Solana queries, retrieving all balance updates for parent transactions * **Network monitoring tool**. ZK-proof based network monitoring tool launched for real-time network health tracking ## Network & Performance * **Noise protocol**. Added to network communication; obfuscates query patterns and prevents stream inspection * **Gzip stream joining**. New gzip stream joining for improved Portal response performance * **Performance-aware query routing**. Query routing now favors workers that meet the published performance requirements and temporarily pauses requests to workers after errors or timeouts. * **Worker benchmark**. A containerized benchmark now checks whether worker hardware meets the Network requirements. * **Portal 503 errors**. Fixed public portal 503 errors * **Sonic Testnet**. Dataset updated * **Monad Mainnet**. Added to Portal ### SDK Updates * **TypeORM refactor**. `createDataSource` renamed to `initializeDataSource` * **Portal client**. Data ingestion and query schemas refactored * **Typegen cleanup**. Removed `solana-stream` dependency from typegen ## New docs site We heard you. Our docs were hard to navigate. The new docs at [docs.sqd.dev](https://docs.sqd.dev) are organized around what you're trying to do. Want to pull raw blockchain data? Start with Portal. Building an indexer? Go straight to the SDK that matches your stack. We also added full documentation for the Pipes SDK. * **Network explorer**. Browse and filter all 100+ supported networks by VM type with real-time status. Plasma, Monad, HyperEVM, and MegaETH added. * **Interactive API playground**. Test [Portal](/en/portal/overview) queries directly in the docs without switching to your terminal. The site is currently in beta while we review and fill in missing content. We plan to leave beta in H1 2026. ### Portal Performance & Features * **Portal consolidation**. Portal trimmed to 11 datasets, all with real-time data: arbitrum-one, avalanche-mainnet, base-mainnet, binance-mainnet, ethereum-mainnet, hyperliquid-mainnet, optimism-mainnet, polygon-mainnet, solana-devnet, solana-mainnet, zora-mainnet * **Zstd compression**. Portal now supports zstd compression (requires workers ≥2.6.0-rc1) * **Performance improvement**. Public portal now delivers ≥8 MB/s of compressed data * **Rate limiting visibility**. Users can now see when Portal bandwidth is being throttled * **`head-number` header**. Portal now returns `head-number` header showing latest available block * **`first_block` metadata**. Metadata now includes `first_block` when no historical data is available * **Dedicated stream endpoints.** [Portal v0.8.1](https://github.com/subsquid/sqd-portal/releases/tag/v0.8.1) fixes `/finalized-stream` to include finalized Hotblocks data and adds `/archival-stream` for Network data only. * **Manual serve mode.** Self-hosted Portal instances can exclude unused datasets. ### Network Worker * **Worker 2.6.0**. [Worker 2.6.0](https://github.com/subsquid/worker-rs/releases/tag/v2.6.0) fixes performance issues, adds zstd compression, corrects Prometheus metric types, and handles missing query files. ### Network App * **Network App Refresh**. Refreshed dashboard design with two new tabs: **Workers** and **Analytics** * **Worker filtering**. Worker filtering (instead of sorting 2,500+ rows) * **Analytics section**. Analytics section with network health, usage, and economics * **Worker analytics**. Each worker page now shows analytics for that worker. ### SDK * **Pipes SDK**. Hyperliquid fills integration, Postgres state management, Drizzle ORM support, factory address handling * **ETHGlobal templates**. Pipes SDK templates created for ETHGlobal hackathon * **Squid SDK improvements**. Etherscan v2 API support in typegen, Solana ESM-compatible typegen import paths, Substrate `SystemOrigin` definition updates, HTTP client 521/522/523 errors now retriable ### New Chains * **[Hyperliquid EVM Mainnet](/en/data/evm/hyperevm-mainnet)** added to Portal * **Beam Mainnet** added to v2 and Portal ## Public Portal GA The [public portal](/en/portal/overview) is now generally available. What started as an internal alpha a year ago is now the default way to access blockchain data through SQD. A [decentralized network of workers](/en/network/overview) serving 100+ datasets with real-time and historical data. The feature-complete portal includes finalized block queries, rate limiting, per-request tracing, and zstd compression. ### Portal & Cloud * **Cloud billing upgrade**. Billing backend upgraded with credits and discounts per [organization](/en/cloud/resources/organizations) ### Network * **Fourteen-day lock periods**. Worker deregistration and delegation withdrawals now each have a 100,000-block lock period, approximately 14 days. ### Pipes SDK * **Pipes SDK documentation**. Learn-by-example documentation published for Pipes SDK ([repo](https://github.com/subsquid-labs/pipes-sdk-docs)) * **Solana instruction decoder**. Added Solana instruction decoder with improved rollback chain logic * **Composite transformer**. Composite transformer with duplicate detection * **HTTP client rework**. HTTP client reworked with improved performance and reliability * **ClickHouse improvements**. Enhanced ClickHouse config and integration * **Fork handling**. Improved fork handling capabilities * **Finalized block queries**. PortalClient now supports querying finalized blocks * **Block hash in decoded events**. Decoded events now include block hash information * **DecodedEvent factory**. Added factory details for DecodedEvent ### Squid SDK * **Portal client fixes**. Token balance schema validation fix, promise chaining memory leak fix * **Solana improvements**. Extend programs missing invoke log, preserve leading `_` in generated names ### New Chains & Hotblocks * **Hedera EVM** added to Portal * **Plasma Testnet and Mainnet** added * **Asset Hub Paseo** added to v2 and Portal * **Solana Devnet hotblocks** enabled * **Binance Mainnet hotblocks** enabled ## SDK & Network Features * **Hasura support**. [Hasura integration](/en/cloud/reference/hasura) released; dedicated Hasura addon for Cloud squids with configuration tool, example repo, and documentation; OpenReader docs consolidated * **Network sharding**. Network protocol now supports sharding for improved scalability * **Network worker 2.5.0**. [Worker 2.5.0](https://github.com/subsquid/worker-rs/releases/tag/v2.5.0) reduces assignment-file size and memory use, supports the updated Solana discriminator syntax, and reports its version in request headers. * **Network worker 2.5.1**. [Worker 2.5.1](https://github.com/subsquid/worker-rs/releases/tag/v2.5.1) stops unnecessary requests to private network addresses. * **Network worker 2.5.2**. [Worker 2.5.2](https://github.com/subsquid/worker-rs/releases/tag/v2.5.2) keeps idle connections open to improve network stability. Worker versions before 2.4.4 are no longer supported. * **Squid generator beta**. New squid generator released in beta; generate squids from YAML configuration * **Solana Devnet**. `solana-devnet` available from block 360,000,000 * **`X-Sqd-Data-Source` header**. Portal responses now include header indicating whether data came from live, hotblocks, or archive * **Pipes SDK foundation**. Solana RPC latency watcher, metrics server, dashboard components, profiler ### Cloud Improvements * **RPC billing fix**. Fixed RPC billing overcharging issue * **Secrets sorting**. Secrets page now supports sorting by name * **Invoice downloads**. Users can now download invoices per month in the Billing section (admin only) ### SDK Improvements * Portal client: `discriminator` field for Solana query, fork error structure fix * Substrate: handle EIP-7702 transactions ## Full Solana History on Portal [Solana's](/en/data/networks/solana) complete mainnet history is now accessible through [Portal](/en/portal/overview). Every block, every transaction, from genesis. Combined with Solana Mainnet hotblocks going live the same month, Portal now covers the full Solana timeline from historical to real-time in a single `solana-mainnet` dataset. ### Portal & Network * **Polygon Mainnet hotblocks** enabled * **Zora Mainnet**. `zora-mainnet` real-time dataset ready, full data from block 0 * **Rewards Claiming App**. App for claiming missing SQD rewards launched at rewards.sqd.ai * **Network worker 2.4.0**. [Worker 2.4.0](https://github.com/subsquid/worker-rs/releases/tag/v2.4.0) adds experimental Hyperliquid support. * **Network worker 2.4.4**. [Worker 2.4.4](https://github.com/subsquid/worker-rs/releases/tag/v2.4.4) uses compact binary assignment files so the Network can assign more data. Query logs now include request IDs. * **Network worker 2.4.5**. [Worker 2.4.5](https://github.com/subsquid/worker-rs/releases/tag/v2.4.5) keeps the assignment stream running after an error. * **Network worker 2.4.6**. [Worker 2.4.6](https://github.com/subsquid/worker-rs/releases/tag/v2.4.6) respects `Retry-After`, avoids aggressive HTTP retries, blocks redirects, and drops empty query messages. * **Starknet RPC ingest**. Added Starknet RPC data ingestion support (since retired) ### SDK * **`@subsquid/portal-client`**. New Portal client package with native Portal integration, response validation, and casting * **Solana Spray client**. Solana Spray client introduced for parallel RPC access ## Portal & Network * **429 rate limiting**. Portal now returns HTTP 429 status when worker CU pool is exhausted; rate limit errors treated as retriable * **Hotblocks status**. Solana hotblocks working in mainnet; EVM hotblocks enabled in testnet, disabled in mainnet (receipts being added) ### SDK * **Irys integration**. Publish datasets to Irys network * **Solana normalization**. Backport new block normalization ## Portal Features * **EVM Hotblocks**. Hotblocks for `ethereum-mainnet` and `base-mainnet` added to public portal * **Solana datasets unified**. Solana datasets reunited into single `solana-mainnet` (historical + hotblocks) * **EVM hotblocks request handling**. Improved request handling for frequently-accessed blocks * **Base hotblocks testing**. Base chain hotblocks available for testing (10,000 last blocks) ### Cloud & Network * **RPC limit raised**. RPC requests spending limit raised from \$200 to \$1,000 * **Hyperliquid onboarded**. [Hyperliquid](/en/portal/hyperliquid/overview) dataset added to the decentralized network ### SDK * **Squid SDK**. Fix name collision in debug frame schema builder ## Portal & Network * **Network scheduling**. New scheduling algorithm deployed to mainnet * **Request-ID tracing**. Per-request identification added to Portal queries for improved debugging and tracing ## SDK & Network * **MegaETH Testnet docs**. MegaETH Testnet documentation page added * **svmBNB & Eclipse Testnet**. SvmBNB Mainnet and Testnet datasets added to Portal testnet; Eclipse Testnet dataset added (starting from block 86,800,000) * **Network worker 2.2.0**. [Worker 2.2.0](https://github.com/subsquid/worker-rs/releases/tag/v2.2.0) replaces gossipsub status exchange with direct status requests. * **Network worker 2.3.1**. [Worker 2.3.1](https://github.com/subsquid/worker-rs/releases/tag/v2.3.1) disables gossipsub by default, reducing worker bandwidth and CPU use. * **Network worker 2.3.2**. [Worker 2.3.2](https://github.com/subsquid/worker-rs/releases/tag/v2.3.2) removes a limit on Portal allocations that could prevent workers from fetching them. ## Portal Open Beta [Portal](/en/portal/overview) is now open to everyone. After three months of closed beta with select Cloud customers, we removed the gates. Any developer can now pull blockchain data through our [decentralized network of workers](/en/network/overview). No waitlist, no approval process. For teams that want full control, Portal is also available as a [self-hosted package](/en/portal/self-hosting). Deploy it on your own infrastructure with docker compose and monitor everything through the bundled Grafana dashboards. ### Migration & Sunset * **Cloud migration path**. Existing Cloud squids can switch from centralized v2 archives to Portal-powered data sources with minimal changes ([EVM/Substrate migration](/en/portal/migration), [Solana migration](/en/sdk/migration/gateway-to-portal#solana)) * **Planned sunset of v2 archives**. With Portal now publicly available, we announced the planned deprecation of centralized v2 archives ### SDK 1.1 Shipped SDK 1.1 is the companion release. It adds native Portal and hotblocks support without breaking existing APIs. If you have a squid running on v2 archives, migration is a config change, not a rewrite. ### Soldexer Brand * **Solana sub-brand**. Solana-focused sub-brand launched with separate landing page and docs ## Cloud & Network * **RPC proxy**. [Cloud RPC proxy](/en/cloud/resources/rpc-proxy) now supports Berachain, Sonic, Unichain, and Unichain Sepolia * **Starknet Sepolia**. Starknet Sepolia added * **Hyperliquid L1 Mainnet**. [Hyperliquid L1 Mainnet](/en/portal/hyperliquid/overview) data available * **Ink Mainnet traces**. Ink Mainnet now supports traces * **Network worker 2.1.0**. [Worker 2.1.0](https://github.com/subsquid/worker-rs/releases/tag/v2.1.0) supports the new Solana dataset format, uses faster compression, and delays assignment switching until Portals can download the new assignment. * **Network worker 2.1.1**. [Worker 2.1.1](https://github.com/subsquid/worker-rs/releases/tag/v2.1.1) limits assignment timeouts to downloads instead of parsing, formats nested query fields as camelCase, and matches addresses case-insensitively. ### New Chains * **Soneium Mainnet** and **Soneium Minato Testnet** added * **Poseidon Testnet** added * **Ink Sepolia** and **Asset Hub Westend Next** added to v2 archives * **Plume** network names and chain IDs swapped ## Network Additions * **Unichain Mainnet**. Unichain Mainnet added to v2 archives * **Portal API specification**. Public [Portal API specification](/en/portal/overview) published * **Berachain Mainnet**. Berachain Mainnet, Monad Testnet, MemeCore Mainnet, and Asset Hub Westend added to v2 archives * **Combined data client**. Combine multiple data streams into one unified client * **Ozean & Hyperliquid Testnets**. Ozean Testnet and Hyperliquid Testnet added to v2 archives * **Lukso Mainnet**. Lukso Mainnet added to v2 archives * **Ink Mainnet, BOB Sepolia, Camp Network Testnet** added to v2 archives * **Bittensor EVM**. Mainnet and testnet added with starting blocks * **Abstract Mainnet & Testnet**. Abstract Mainnet and Testnet added to v2 archives * **Soon Mainnet**. Soon Mainnet (Solana-based) added to v2 archives ## Portal Closed Beta This is where [Portal](/en/portal/overview) started. After months of internal development, we shipped the first version of our decentralized data gateway to a closed group of [Cloud](/en/cloud/overview) customers. The goal: prove that a [decentralized network of workers](/en/network/overview) could serve production blockchain data as reliably as our centralized archives. ### Launch Features * **Portal web console**. A new UI for browsing datasets, testing queries, and monitoring Portal usage * **Portal-powered Cloud squids**. Existing Cloud squids could now run on Portal instead of centralized archives, with no code changes required * **\$100 migration credit**. Early migrators received a \$100 credit to incentivize switching from v2 archives to Portal * **Substrate support**. [Substrate](/en/data/networks/substrate) chain data added to Portal alongside [EVM](/en/data/networks/evm) and [Solana](/en/data/networks/solana) * **New SQD.ai website**. Redesigned company website launched at sqd.dev ## Network worker 2.0.0 [Network worker](/en/network/worker) 2.0.0 changed worker communication to use compact messages, rate limiting, reliable log delivery under load, and signed query results. Worker configuration moved to a single `.env` file. ## Cloud Deployments 2.0 [Cloud](/en/cloud/overview) Deployments got a major overhaul. The old model. One squid, one version, one endpoint. Didn't scale for teams running multiple environments or doing zero-downtime upgrades. Deployments 2.0 introduces **[slots](/en/cloud/resources/slots-and-tags)** (isolated execution environments) and **tags** (movable pointers like `production` or `staging`) so you can deploy, test, and promote without touching your live endpoint. ([release notes](https://docs.sqd.dev/deployments-two-release-notes/)) ### Cloud & SDK Updates * **Squid CLI 3.0.0**. CLI overhauled to match: commands renamed for clarity (`rm`→`remove`, `ls`→`list`), new deployment flow with tagging, confirmation prompts for destructive actions * **[Tron](/en/data/tron/tron-mainnet) support**. All Tron-related npm packages published, with [TRC-20 transfer tracking example](https://github.com/subsquid-labs/tron-example) * **Network worker v1.1.1**. [Network worker](/en/network/worker) v1.1.1 released on mainnet * **L1 fee fields**. EVM processor now supports L1 fee fields on Optimism for financial reporting ### New Chains * **Mode Mainnet** and **Degen Chain** added * **Canto Testnet** added * **Aleph Zero EVM**, **Berachain bArtio Testnet**, **StratoVM Sepolia**, **Arthera Mainnet** added * **B3 network** added ## SDK & Cloud Features * **Polygon traces**. Polygon Mainnet traces support went live * **SDK examples**. New curated examples page published in docs * **Fuel SDK 1.1.0**. Fuel data packages updated: block consistency validation with retry, optional block transactions * **Multiple API tokens**. Cloud API now supports multiple API tokens per account * **ink! v5 support**. Ink! v5 smart contracts now supported in Substrate SDK ### New Chains * **NeoX Testnet**, **B3 Sepolia**, **Plume Testnet**, **Galxe Gravity (EVM)**, **Gemini 3h**, **Bittensor Testnet (Substrate)** added; Watr network removed * **0g Testnet** added * **Acurast Canary (Substrate)** added * **Zora Sepolia**, **Kyoto Testnet**, **Dogechain Testnet**, **Etherlink Mainnet**, **BOB Mainnet** added * **Taiko Mainnet (EVM)** and **Eclipse Testnet (Solana, beta)** added ## Network worker 1.0.0-rc2 [Worker 1.0.0-rc2](https://github.com/subsquid/worker-rs/releases/tag/v1.0.0-rc2) reduces RAM use, adds timeouts for stalled chunk downloads, and broadcasts query logs instead of sending them directly. ## June 2026 * **Berlin Blockchain Week + Telekom meetup**. SQD at Berlin Blockchain Week (June 13–21) and Telekom & Friends Vol. 3: Blocks & Brews, a Deutsche Telekom networking meetup at hubraum on June 18, continuing the partnership with Telekom Web3 Infrastructure & Solutions * **AI-native data pipelines at Blockchain Community Day 2026**. Product Lead [@karelxfi](https://x.com/karelxfi) on feeding raw blockchain data to autonomous systems without killing performance ([free event, 3pm CET](https://luma.com/wupxaqku)) * **Turtle × SQD**. Turtle ([@turtledotxyz](https://x.com/turtledotxyz)), a Liquidity Distribution Protocol for internet capital markets, uses SQD Portal to power its data layer, so LPs find the best deals and asset issuers find LPs ## May 2026 * **SQD Stream Podcast (Ep. 1)**. First episode of the SQD Stream podcast ([watch](https://sqd.dev/videos/sqd-stream-podcast-ep-1/)) * **What is SQD?** Explainer clip from the SQD Stream podcast with [@kureus\_](https://x.com/kureus_), all of Web3's data in one place via a single scalable API ([watch](https://sqd.dev/videos/sqd-stream-podcast-ep-1/)) * **Long live the SQD!** Behind-the-scenes look at the SQD team and the culture building SQD Portal ([video](https://sqd.dev/videos/long-live-the-sqd/)) * **AI agent usage up \~200%**. AI agent usage on SQD Portal up about 200%, including an agent pairing SQD Portal with DeepSeek for Hyperliquid data more complete than UI scraping * **SQD Buy & Burn explained**. SQD Buy & Burn, powered by enterprise data demand, with a live [Portal Pools dashboard](https://network.sqd.dev/dashboard/portal-pools) ## April 2026 * **Snoopy beta**. In-depth piece on block header verification at ingestion plus a preview of Snoopy, an experimental on-chain fraud proofs and verification layer by engineer Denis. Snoopy beta is now live ([announcement](https://x.com/helloSQD/status/2046637866890252550)) * **240+ TB served, 85M+ queries in two weeks**. Polygon led on data served, Arbitrum led on queries ([announcement](https://x.com/helloSQD/status/2046952244944183686)) * **Istanbul offsite**. SQD team offsite in Istanbul, "entering the next phase" ([announcement](https://x.com/helloSQD/status/2048803492114878584)) * **DK at Slovak University of Technology**. Core engineer DK presented SQD Network at FIIT STU ([announcement](https://x.com/helloSQD/status/2049519110371901665)) * **SQD Stream Podcast Ep. 1**. CPO and Mojo discuss Web3 data problems, the enterprise roadmap, and Revenue Pools, recorded during the Istanbul offsite ([video](https://x.com/helloSQD/status/2049894508154716198)) ## March 2026 * **Lambda case study**. Lambda chose SQD for multi-chain portfolio data ([part 1](https://x.com/helloSQD/status/2033632513499795957?s=20)) * **Katana added to SQD Network**. New chain support ([announcement](https://x.com/helloSQD/status/2029968722727764447?s=20)) * **AI development tools in docs**. AI development tooling section added to [SQD Docs](/en/ai/ai-development) ## February 2026 * **SQD Q1 update**. Quarterly progress update ([announcement](https://x.com/helloSQD/status/2025978234865942802?s=20)) * **ETHDenver Infra Leaders Brunch**. SQD hosted an infrastructure leaders brunch at ETHDenver ([video](https://x.com/helloSQD/status/2024863740915806418?s=20)) * **ADI Mainnet indexing**. SQD to provide indexing infrastructure for ADI Mainnet ([announcement](https://x.com/helloSQD/status/2024529663461114312?s=20)) * **18M queries in 24 hours**. SQD Network milestone ([announcement](https://x.com/helloSQD/status/2023441306643485129?s=20)) * **Solana Data Room**. Interactive Solana network statistics dashboard launched at [room.soldexer.dev](https://room.soldexer.dev/feb-2026), powered by SQD ([announcement](https://x.com/helloSQD/status/2021625892913787388?s=20)) * **zkVerify case study**. ZkVerify uses SQD for real-time indexing and cross-chain monitoring of ZK proof data ([announcement](https://x.com/helloSQD/status/2020900250098266478?s=20)) * **Portal 45x faster**. Portal indexing upgrade delivers 45x faster performance ([announcement](https://x.com/helloSQD/status/2019828389389992303?s=20)) * **Lambda on SQD**. Solana DeFi platform Lambda goes live on SQD ([announcement](https://x.com/helloSQD/status/2019108221215207904?s=20)) * **Polygains**. Community member launched [Polygains](https://polygains.com/), a MEV analytics platform built with SQD ([announcement](https://x.com/helloSQD/status/2023802079769489905), [source code](https://github.com/mev-tools/polygains/)) * **Agent Skills on skills.sh**. SQD agent skills published to [skills.sh](https://skills.sh/?q=subsquid) * **Agentic Commerce Market Map**. SQD featured in [Agentic Commerce Market Map](https://x.com/helloSQD/status/2027079167880396966) as blockchain data infrastructure for AI commerce agents ## January 2026 * **Revenue Pools explainer**. Deep dive into sustainable rewards powered by demand ([announcement](https://x.com/helloSQD/status/2017284724398428313?s=20)) * **Revenue Pools LIVE**. Revenue Pools launched ([announcement](https://x.com/helloSQD/status/2014037225982345694?s=20)) * **Revenue Pool oversubscribed**. First Revenue Pool oversubscribed ([announcement](https://x.com/helloSQD/status/2011144916777279811?s=20)) * **Curvance case study**. Curvance cut indexing time from days to minutes with SQD ([blog](https://blog.sqd.dev/curvance-cut-indexing-time-from-days-to-minutes-with-sqd/)) * **Revenue Pools flow diagram**. Visual explainer of Revenue Pools mechanics ([announcement](https://x.com/helloSQD/status/2008945817135272375?s=20)) * **Revenue Pools on Binance**. Revenue Pools coverage on Binance Square ([article](https://www.binance.com/en-GB/square/post/34447045316321)) * **Year in review**. SQD annual recap ([announcement](https://x.com/helloSQD/status/2006744577362002412?s=20)) ## December 2025 * **Revenue Pools announced**. SQD launches Revenue Pools backed by enterprise customer payments ([blog](https://blog.sqd.dev/sqd-launches-revenue-pools-backed-by-enterprise-customer-payments/), [HackerNoon coverage](https://hackernoon.com/sqd-network-just-killed-token-emissions-heres-what-\$16-billion-in-defi-tvl-pays-instead)) * **CPO on Rollup TV (Part 2)**. Dan shares insights on e-commerce, AI agents, and stablecoins ([video](https://x.com/helloSQD/status/2000667684912964018?s=20)) * **Pipes SDK walkthrough**. Video walkthrough of the Pipes SDK ([video](https://x.com/helloSQD/status/1999191474018382065?s=20)) * **Blockster coverage**. Blockster covers SQD Portal and animated explainer ([article](https://blockster.com/subsquids-cultural-breakthrough-carbot-animations-brings-the-portal-to-life)) * **Portal animated explainer**. SQD Portal: The Movie ([video](https://x.com/helloSQD/status/1998433109704880247?s=20)) * **Nodies one-click staking**. Staking walkthrough with Nodies ([blog](https://blog.sqd.dev/nodies-one-click-staking-for-sqd/)) * **ETHGlobal Buenos Aires hackathon**. Hackathon highlights and recap ([blog](https://blog.sqd.dev/sqd-hackathon-at-ethglobal-buenos-aires/), [video](https://x.com/helloSQD/status/1996589853417283767?s=20)) * **SQD Docs 2.0 (beta)**. New documentation site launched at [docs.sqd.ai](https://docs.sqd.ai/) * **November recap**. Monthly highlights including 7,500 unique delegators, ETHGlobal, Gate Alpha listing, Portal beta launch ([recap](https://x.com/helloSQD/status/1995535932405993970?s=20)) ## November 2025 * **856 TB served**. 856 TB served in 90 days, 11 TB in the last 24h ([milestone](https://x.com/helloSQD/status/1994458776162693127?s=20)) * **x402 x SQD x Rezolve Ai**. AI agents and the future of e-commerce ([blog](https://blog.sqd.dev/agents-can-pay-now-let-them-think-x402-x-sqd-x-rezolve-ai/)) * **Monad ecosystem support**. SQD powering Monad builders from testnet to mainnet ([announcement](https://x.com/helloSQD/status/1993043652847837236?s=20)) * **Blockchainreporter coverage**. Media coverage of SQD network upgrade and streaming data through Portal ([article](https://blockchainreporter.net/sqds-network-upgrade-will-deliver-streaming-data-through-portal-heres-why-its-a-big-deal/)) * **ETHGlobal hackathon winners**. Data Context Market won \$3.5k at ETHGlobal ([announcement](https://x.com/helloSQD/status/1992731999128400286?s=20)) * **ETHGlobal Buenos Aires**. SQD at ETHGlobal event ([highlights](https://x.com/helloSQD/status/1991922796310720563?s=20), [recap](https://x.com/helloSQD/status/1992273492725862671?s=20)) * **Product philosophy blog**. Portal, Pipes SDK, and the product philosophy behind SQD ([blog](https://blog.sqd.dev/portal-pipes-sdk-and-the-product-philosophy-behind-sqd/)) * **Data Moat Interviews (Ep. 3)**. Ops Lead Thomas on decentralizing data ([video](https://x.com/helloSQD/status/1989004309258818001?s=20)) * **CPO on Rollup TV**. Dan Quirk on AI in 2025 and SQD's endgame vision ([video](https://x.com/therollupco/status/1988684798328734177?s=20)) * **SQD Network explained**. Core engineer DK explains SQD Network ([blog](https://blog.sqd.dev/sqd-network-explained-by-dk/)) * **Network Dashboard upgrade**. Live views of Network Health, Economics, and Usage ([announcement](https://x.com/helloSQD/status/1987931489728688340?s=20)) * **Data Moat Interviews (Ep. 2)**. CPO Dan on Portal and the next 12 months for SQD ([video](https://x.com/helloSQD/status/1986500272643015100?s=20)) * **Portal beta + 5.8M daily queries**. Portal beta rollout milestone ([announcement](https://x.com/helloSQD/status/1985780453844955510?s=20)) ## October 2025 * **Data Moat Interviews (Ep. 1)**. DZ on the vision behind SQD ([video](https://x.com/helloSQD/status/1983948189238452660?s=20)) * **5 Reasons Portal is the new standard**. Product deep dive on Portal ([blog](https://blog.sqd.dev/5-reasons-portal-is-the-new-standard-for-blockchain-data/)) * **SQD on Bybit**. Token listing on Bybit ([announcement](https://x.com/helloSQD/status/1978166405208899946?s=20)) * **5.33 million daily queries**. Usage milestone ([announcement](https://x.com/helloSQD/status/1977785442398253547?s=20)) * **Rezolve AI acquires Subsquid**. Building the data and payments backbone for the AI economy ([blog](https://blog.sqd.dev/rezolve-ai-acquires-subsquid-sqd/), [GlobeNewswire](https://globenewswire.com/news-release/2025/10/09/3164101/0/en/Rezolve-Ai-Acquires-Subsquid-Building-the-Data-and-Payments-Backbone-for-the-AI-Economy.html), [Decrypt](https://decrypt.co/343728/rezolve-ai-acquires-subsquid-building-the-data-and-payments-backbone-for-the-ai-economy), [Bitcoin.com](https://news.bitcoin.com/rezolve-ai-acquires-sqd-to-power-web3-driven-enterprise-ai), [Investing.com](https://investing.com/news/company-news/rezolve-acquires-blockchain-data-platform-subsquid-in-ai-commerce-push-93CH-4279417), [Blockmedia KR](https://blockmedia.co.kr/archives/989486)) * **3,000 worker nodes online**. Infrastructure scaling milestone ([announcement](https://x.com/helloSQD/status/1973754542241263927?s=20)) * **Public Portal GA**. The [public portal](/en/portal/overview) is now generally available, serving 100+ datasets with real-time and historical data ## September 2025 * **OceanStream: Open Token Facts**. Listing attestation framework with schemas and attestations ([blog](https://blog.sqd.dev/oceanstrea/)) * **Chillwhales case study**. NFT and entertainment integration with SQD ([blog](https://blog.sqd.dev/chillwhales-x-sqd-partner-case-study/)) * **From the SEC Roundtable to Real-Time Markets**. Trust and data transparency focus ([blog](https://blog.sqd.dev/from-the-sec-roundtable-to-real-time-markets-how-trust-unlocks-the-onchain-economy/)) * **SQD Interview with DZ**. Interview with SQD's DZ ([video](https://x.com/helloSQD/status/1966549332796055699?s=20)) * **Full Solana history**. Entire Solana history available from genesis ([announcement](https://x.com/helloSQD/status/1965471408588275722?s=20)) * **OceanStream at the SEC Crypto Taskforce Roundtable**. SQD's institutional engagement ([blog](https://blog.sqd.dev/oceanstream-at-the-sec-crypto-taskforce-roundtable/)) * **Levr.Bet case study**. On-chain sports data use case ([blog](https://blog.sqd.dev/levr-bet-x-sqd-partner-case-study/)) ## August 2025 * **CoolWallet case study**. How CoolWallet uses SQD to secure its data back-end ([blog](https://blog.sqd.dev/how-coolwallet-uses-sqd-to-secure-its-data-back-end/)) * **Monthly recap: July**. Tech and community updates ([blog](https://blog.sqd.dev/monthly-recap-july/)) * **OceanStream positioning**. OceanStream as the institutional arm of SQD ([announcement](https://x.com/helloSQD/status/1952330964060749864?s=20)) ## July 2025 * **SQD in the Ethereum Ecosystem**. Integration with Ethereum apps ([blog](https://blog.sqd.dev/sqd-in-the-ethereum-ecosystem/)) * **Global permissionless database vision**. Long-term vision for transcending data silos ([announcement](https://x.com/helloSQD/status/1947325708264788421?s=20)) * **The Story of SQD**. Founding story, mission, and values ([blog](https://blog.sqd.dev/the-story-of-sqd/)) * **What happened at SQD in June**. Monthly milestone recap ([blog](https://blog.sqd.dev/what-happened-at-sqd-in-june/)) * **SQD in Numbers**. Key metrics on query volume, TVL growth, and nodes ([blog](https://blog.sqd.dev/sqd-in-numbers)) ## June 2025 * **Rose AI x OceanStream partnership**. Strategic partnership for institutional on-chain data access ([Blockmedia KR](https://blockmedia.co.kr/archives/936261)) * **Uniswap V4 data with the Squid SDK**. How to index and query Uniswap V4 data using the Squid SDK ([blog](https://blog.sqd.dev/unlock-uniswap-v4-data-in-minutes-with-sqd-sdk)) * **Coinbase listing**. SQD token goes live on Coinbase on the Arbitrum network ([announcement](https://x.com/helloSQD/status/1932838399708442804?s=20)) * **Vision teaser**. SQD working on changing the database space ([announcement](https://x.com/helloSQD/status/1929526044622717404?s=20)) ## May 2025 * **SQD Data Talk**. Launch of a technical series with a deep dive into RPC proxy design in Web3 ([blog](https://blog.sqd.dev/sqd-data-talk-insights-on-building-an-rpc-proxy)) ## April 2025 * **SQD x io.net partnership**. Partnership for high-performance Web3 and AI infrastructure ([Blockmedia KR](https://blockmedia.co.kr/archives/883234)) ## Portal Open Beta [Portal](/en/portal/overview) is now open to everyone. After three months of closed beta with select Cloud customers, we removed the gates. Any developer can now pull blockchain data through our [decentralized network of workers](/en/network/overview). No waitlist, no approval process. * **Soldexer brand**. Solana-focused sub-brand launched with separate landing page and docs ## Portal Closed Beta This is where [Portal](/en/portal/overview) started. After months of internal development, we shipped the first version of our decentralized data gateway to a closed group of [Cloud](/en/cloud/overview) customers. * **New SQD.ai website**. Redesigned company website launched at sqd.dev ## Related topics - [Changelog: slots and tags](/en/cloud/reference/deployments-two-changelog.md) - [Slots and tags](/en/cloud/resources/slots-and-tags.md) - [Production Deployment](/en/cloud/reference/cli/prod.md) - [Developer resources](/developers.md) - [Adding API keys to gateways](/announcements/gateway-api-keys.md) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.