DocSearch MCP
DocSearch MCP lets AI clients search current public developer documentation from the DocSearch corpus.
Use it when you want an assistant to answer questions from public docs instead of relying only on model training data. The public endpoint does not require authentication:
https://mcp.algolia.com/1/docsearch/mcp
What it does​
DocSearch MCP exposes documentation search through the Model Context Protocol. MCP-compatible clients connect to the endpoint and call DocSearch tools while answering your questions.
The endpoint is focused on public developer documentation. You do not need an Algolia application ID, search API key, or DocSearch application to use it.
How it works​
Most lookups are a single call: name the product and ask your question, and DocSearch finds the right documentation set and returns the matching content together.
When a question spans several products, or you want to inspect and hand-pick documentation sets first, there is a two-step flow: resolve the documentation sets, then query the ones you choose.
Available tools​
algolia_docsearch_search_docs​
The one-shot tool, and the right default for most lookups. Give it a library (the product, SDK, or platform) and a query (your question); it resolves the best matching documentation set and returns ranked content in a single call. If the library is ambiguous, it returns candidate documentation sets to choose from instead.
algolia_docsearch_resolve_docset​
Step 1 of the manual flow. Finds the documentation sets that best match a product, library, or platform and returns candidates — each with a docset_id, title, description, and ranking signals to help pick the best match.
algolia_docsearch_query_docs​
Step 2 of the manual flow. Retrieves documentation content for one or more docset_ids returned by algolia_docsearch_resolve_docset. Pass several at once when a question spans multiple products.