MCP SERVER · OFFICIAL PROJECTPinecone's developer server — search the docs, configure indexes, and upsert and query records — so a coding assistant works against your real Pinecone project.
LABSCO SUMMARY
search-docs is the point of this being Pinecone's own server: the assistant grounds itself in the current documentation before generating index config or query code, rather than relying on what a model happens to remember. The upsert and search tools then let it test those queries against your real data in the dev loop — but note they only work on integrated-inference indexes, so describe-index is the check to run first.
Pinecone's official developer server, aimed at coding assistants: 9 tools for searching Pinecone's documentation, discovering and creating indexes, and upserting and searching records against your project. Runs over stdio.
Node.js v20 or later with node and npx on PATH, run as npx @pinecone-database/mcp. A Pinecone API key in PINECONE_API_KEY from app.pinecone.io — without it the assistant can still search the documentation, but cannot manage or query your indexes. The search and upsert tools work on integrated-inference indexes, which embed text automatically; describe-index tells you an index's type.
One command plus a key — npx @pinecone-database/mcp, then supply credentials
USE CASES
Search the official documentation for the behavior in question, then generate code informed by it rather than by assumptions the model may have picked up.
Create an integrated-inference index that embeds your text field automatically, waiting only for the call to return before you start loading data.
Upsert records into the index, confirm they landed by checking per-namespace stats, then run a semantic search and rerank the hits to evaluate the results.
Run a cascading search over multiple integrated indexes, letting it deduplicate and rerank the combined results into one ranked list.
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
cascading-searchSearch across multiple indexes for records that are similar to the query text, then deduplicate and rerank the combined results. Reranking is required for this tool. Only works with integrated-inference indexes. To search a single index, use search-records instead.
create-index-for-modelCreate a Pinecone index with integrated inference, which automatically embeds the text field named in "fieldMap". Supports AWS, GCP, and Azure cloud providers. This call blocks until the index is ready to accept upserts and queries, so no readiness polling is needed afterwards. If an index with the same name already exists, it is returned unchanged instead of erroring.
describe-indexDescribe the configuration of a Pinecone index, including its embedding model and the "fieldMap" that names the record field holding the text to embed. Call this before upsert-records to learn the required field name. Check "status.ready" in the response to confirm the index can accept upserts and queries.
describe-index-statsDescribe the statistics of a Pinecone index, including record counts per namespace. Use this to discover which namespaces exist in an index. Newly upserted records can take a few seconds to be indexed and become visible to search results and index stats; if recent data is missing, wait briefly and retry.
list-indexesList all Pinecone indexes in the project, including each index's configuration and readiness status. Use this to discover valid index names before calling other tools.
rerank-documentsRerank a set of provided documents by relevance to a query. Use this for documents that did not come from a Pinecone search (e.g. externally sourced text). To rerank the results of a single-index search, prefer the "rerank" parameter of search-records instead of calling this tool separately.
search-docsSearch the official Pinecone documentation for relevant information. Use this before writing Pinecone code or explaining Pinecone behavior, rather than relying on assumptions.
search-recordsSearch an index for records that are semantically similar to the query text. Only works with integrated-inference indexes (e.g. created with create-index-for-model), which embed the query automatically; it fails on standard indexes — use describe-index to check the index type. Returns hits with "_id", "_score", and the record's stored "fields". Newly upserted records can take a few seconds to be indexed and become visible to search results and index stats; if recent data is missing, wait briefly and retry.
upsert-recordsInsert or update records in an integrated Pinecone index. Text in the field named by the index's "fieldMap" is embedded automatically — call describe-index first to confirm that field name. Records with an existing "id" are overwritten. Newly upserted records can take a few seconds to be indexed and become visible to search results and index stats; if recent data is missing, wait briefly and retry.
BEFORE YOU INSTALL
9 tools when we connected. Without PINECONE_API_KEY only search-docs works. Only indexes with integrated inference are supported, and Pinecone Assistants need the separate Assistant MCP server.
CONTEXT COST
Rated by the size of the 9 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
npx @pinecone-database/mcp{
"mcpServers": {
"pinecone": {
"command": "npx",
"args": [
"-y",
"@pinecone-database/mcp"
],
"env": {
"PINECONE_API_KEY": "<YOUR_PINECONE_API_KEY>"
}
}
}
}gemini extensions install https://github.com/pinecone-io/pinecone-mcpMCP servers
akshay2211
tomohiro-owada
dodopayments
jsbroks