Artificial intelligence news.
Straight from the original announcements.

Weaviate Deprecates GraphQL API, Cloud Shutdown Set for May 1, 2027

October 26, 2026Source: Weaviate
Weaviate Deprecates GraphQL API, Cloud Shutdown Set for May 1, 2027
Image: Weaviate

Weaviate has deprecated GraphQL, the first query API it offered, and will switch it off on Weaviate Cloud on May 1, 2027. According to the company, nothing breaks on the deprecation date of October 26, 2026, and users have gRPC-based clients and a REST Search API to move to.

Key points
  • Weaviate deprecated its GraphQL API on October 26, 2026, and nothing breaks that day.
  • Weaviate Cloud switches GraphQL off on idle clusters on December 1, 2026 and on all clusters on May 1, 2027.
  • Recommended replacements are gRPC-based clients, the TypeScript web client, and the REST Search API, which is in Preview.
  • Self-hosted versions released before May 1, 2027 keep GraphQL; the first minor release after that disables it by default.
  • Users can test with GraphQL disabled via the DISABLE_GRAPHQL setting or by asking Cloud support.

Timeline for Weaviate Cloud and self-hosted setups

Weaviate says the GraphQL API is deprecated as of October 26, 2026, but that existing clusters keep working for now. On Weaviate Cloud, GraphQL is switched off in two steps. On December 1, 2026, it is turned off on clusters that have not used GraphQL in the last 30 days. On May 1, 2027, which is the end-of-life date for GraphQL on Weaviate Cloud, it is switched off on all remaining clusters.

New Weaviate Cloud clusters are already created without GraphQL, according to the company. Weaviate is emailing all Weaviate Cloud customers so they know whether their clusters are affected.

For self-hosted users, every Weaviate version released before May 1, 2027 keeps a working GraphQL API. The first minor release after that date disables it by default, and it can be switched back on. It will be removed in a later release, for which Weaviate has not set a date yet.

Why Weaviate is moving away from GraphQL

Weaviate says it wants one modern query API and names several reasons.

  • Speed: gRPC sends binary Protocol Buffers over HTTP/2, so queries use less bandwidth and return faster than JSON over HTTP, according to the company.
  • Clear contracts: typed messages give clients typed results and generated request objects. Errors arrive as real status codes rather than as HTTP 200 with an errors array, and metadata is a field on the result instead of a separate _additional block.
  • Security: GraphQL introspection needs read permission on every collection, which Weaviate says does not fit role-based access control (RBAC) with per-collection permissions. The gRPC and REST APIs check exactly the collection being queried.
  • Modern query writing: Weaviate notes that most queries are now built by client libraries, frameworks and agents, which a compact binary protocol and a plain JSON API serve better than a hand-written query language.

Maintaining one API also means each search feature is built, tested and documented once, which the company says lets it ship improvements faster.

Who is affected and what to use instead

The change only affects users who send queries to Weaviate over GraphQL. That includes code that calls /v1/graphql directly, uses raw helpers such as graphql_raw_query() in Python, or posts GraphQL queries from browser code. It also covers older clients: the Python v3 API (weaviate.Client), weaviate-ts-client v2, the Java client v5 and earlier, and the Go client. The stable Go v5 client still queries over GraphQL, while the v6 client uses gRPC and is a release candidate. No-code or BI tools that send GraphQL are affected too. Users of frameworks such as LangChain or LlamaIndex should use the official Weaviate integration built on a current client, since a community package may still send GraphQL.

Weaviate says anyone searching through a current Python, TypeScript, Java or C# client is most likely not affected. It recommends these clients, which search over gRPC. For browsers and edge runtimes that cannot open a native gRPC connection, the TypeScript web client (@weaviate/web) uses gRPC-Web and is available now with the same API as the Node client.

For stacks that cannot run a client, a REST Search API offers plain JSON endpoints for search and aggregation. It is in Preview and is a subset of what GraphQL offers. Weaviate acknowledges that a few GraphQL-only features have no equivalent today and asks users who rely on them to contact support.

How to migrate and what it means for users

Weaviate recommends four steps: find GraphQL calls in code, proxies and tools; pick a replacement path; convert queries using the migration guide; and verify the application with GraphQL disabled. The company's guide includes a decision table and an example of one near-text search on a Movie collection written in GraphQL, the Python client, the TypeScript web client and the REST Search API.

Self-hosted users can test now by setting the DISABLE_GRAPHQL environment variable to true and restarting, or by using the runtime-config key disable_graphql. With GraphQL disabled, POST /v1/graphql and POST /v1/graphql/batch return HTTP 422. On Weaviate Cloud, users can ask support to switch GraphQL off on a cluster of their choice for testing.

With six months between deprecation and end of life on Weaviate Cloud, teams can plan the move around their own releases. Support is available through the support portal or at [email protected].

Related news

Google AI Studio Adds Maps Tools for Prompt-Based App Building
Products and toolsNovember 9, 2026

Google AI Studio Adds Maps Tools for Prompt-Based App Building

Google AI Studio now lets users build apps powered by Google Maps Platform by describing what they want to Gemini. Google says the feature can create tools such as store…

Products and tools
Products and toolsNovember 6, 2026, 12:09

LenDb.ai Launches Free Database of 224,703 US Loan Officers

Draper Goren Blockchain Studio has launched LenDb.ai, a free, lender-neutral database designed to help ChatGPT, Claude and other AI assistants identify US mortgage loan…

Weaviate adds PathSeer, an adaptive filtered vector search, in version 1.40
For developersOctober 14, 2026

Weaviate adds PathSeer, an adaptive filtered vector search, in version 1.40

Weaviate has added PathSeer, a predicate-agnostic filtered vector search algorithm for HNSW, to its database in version 1.40. According to Weaviate, benchmarks show it…