Integration · Company data

Connect Claude & ChatGPT to Pinecone

Ask your AI to search a Pinecone index, look at what is stored in a namespace, or add new records. It can run similarity queries, fetch vectors by ID or metadata, upsert vectors, records and documents, and read index and namespace statistics. You can check and work with your vector data from a chat instead of writing scripts.

Connect Pinecone free

Free plan, no credit card. Takes about a minute.

Try asking

“How many vectors are in the index, and how are they spread across namespaces?”
The AI reads the index statistics: total vector count, dimension and the count per namespace.
“Search the support-docs namespace for passages about refund policy.”
It searches records by meaning with your text as the query and returns the top matches with their fields.
“Find the 10 vectors most similar to vector doc-4812.”
It runs a query with that vector ID, returns the closest matches with scores and, if you ask, their metadata.

What your AI can do

Search and query

Find similar vectors by vector or ID, search records by meaning with optional reranking, and search documents with a metadata filter.

Write and update data

Upsert vectors, records and documents, update vector values or metadata, and partly update documents.

Namespaces and stats

List and create namespaces, list vector and document IDs, and read index statistics per namespace.

“Show me the metadata for vectors faq-001 and faq-002.”
It fetches the two vectors by ID and shows their values and metadata.
“Mark every vector with source=old-site as archived.”
It updates the metadata of all vectors matching the filter, in the namespace you name.
“Remove all vectors from the test namespace.”
It deletes every vector in that namespace. The deletion is permanent and cannot be undone.

18 tools for Pinecone

These tools are switched on when you connect. You can switch any of them off, or require your approval before it runs.

create_namespaceChanges dataCreate a new namespace in the index with a `name`, optionally with a `schema` saying which metadata fields are filterable.
delete_vectorsDeletesPermanently delete vectors from a `namespace` by `ids` or by metadata `filter`; `delete_all` removes every vector in the namespace. This cannot be undone.
describe_index_statsChanges dataGet statistics for the index: total vector count, dimension and per-namespace counts. Optionally pass a metadata `filter`.
fetch_documentsChanges dataFetch documents from a `namespace` by `ids` or by metadata `filter`, choosing which `include_fields` to return.
fetch_vectorsFetch vectors by their `ids` from an optional `namespace`, returning values and metadata.
fetch_vectors_by_metadataChanges dataFetch vectors whose metadata matches a `filter`, in an optional `namespace`. Supports `limit` and `pagination_token`.
get_namespaceGet details for one namespace by `namespace` name, such as its record count.
list_documentsChanges dataList document IDs in a `namespace`, optionally filtered by ID `prefix`. Supports `limit` and `pagination_token`.
list_namespacesList the namespaces in the connected index, optionally filtered by name `prefix`. Use to find which namespace holds the data you need.
list_vector_idsList vector IDs in a `namespace`, optionally only those starting with an ID `prefix`. Supports `limit` and `pagination_token`.
query_vectorsChanges dataFind the `top_k` most similar vectors to a query `vector` or an existing vector `id`, in an optional `namespace` and with an optional metadata `filter`. Set `include_metadata` to return metadata.
search_documentsChanges dataSearch documents in a `namespace` and return the `top_k` best matches. `score_by` lists the scoring methods, e.g. [{"type": "text", "field": "body", "query": "refund policy"}]; add an optional metadata `filter`.
search_recordsChanges dataSearch records in a `namespace` by meaning. Pass `query` with `top_k` and either text `inputs` (for indexes with integrated embedding), a record `id` or a `vector`, plus optional metadata `filter`, `fields` to return and `rerank`.
update_documentsChanges dataPartly update documents in a `namespace`: either pass `documents` (each with `_id` and the fields to change) or a metadata `filter` with `set_fields` / `remove_fields`.
update_vectorChanges dataUpdate one vector by `id` (new `values` or `set_metadata`), or update the metadata of all vectors matching a `filter`, in an optional `namespace`.
upsert_documentsChanges dataAdd or overwrite up to 1,000 `documents` in a `namespace`. Each document needs an `_id`; other fields are stored and validated against the index schema.
upsert_recordsChanges dataAdd or overwrite records in a `namespace` of an index with integrated embedding. Each item in `items` needs an `_id` plus its text and metadata fields; existing records with the same ID are replaced.
upsert_vectorsChanges dataAdd or overwrite vectors in a `namespace`. Each item in `vectors` has an `id`, `values` and optional `metadata`; existing vectors with the same ID are replaced.

Set up in three steps

1
Pick the app
Create a free PipMCP account and choose Pinecone from the app list.
2
Paste your key
Paste two values: your index host and an API key. In the Pinecone console (app.pinecone.io) open your project, go to API keys and click Create API key. Then open Database > Indexes, pick the index and copy its host without https:// (for example my-index-abc123.svc.aped-4627-b74a.pinecone.io). One connection works with one index.
3
Add the link to your AI
You get a personal MCP link. Add it to Claude, ChatGPT or Cursor:
1. Click your name, then Settings › Connectors › Add custom connector.
2. Paste your link as the Remote MCP server URL.
3. Switch it on from the + menu in a chat.

Questions

What can the AI do in Pinecone?

The AI can read index statistics, list and inspect namespaces, list vector and document IDs, fetch vectors and documents, and run similarity queries and searches. It can create namespaces, upsert vectors, records and documents, update vectors and documents, and delete vectors if you switch that tool on.

Does the AI see my Pinecone API key?

No. Your Pinecone API key is encrypted at rest and never shown to the AI. After you save it, it is not shown again, not even to you. The AI only sees the results of the tools it calls.

Can I control what the AI is allowed to do?

Yes. You choose which tools are switched on, so you can start with search and fetch only. Deleting vectors is permanent, and upserts replace existing vectors with the same ID, so those tools can require your approval before they run. Every tool call is logged.

Does it work with ChatGPT?

Yes. In ChatGPT go to Settings › Apps & Connectors › Advanced and turn on Developer mode, then add your PipMCP link. Developer mode needs a paid ChatGPT plan: Plus, Pro, Business or Enterprise. The same link also works in Claude (Settings › Connectors › Add custom connector), Cursor and other MCP clients.

What does it cost?

PipMCP has a free plan with no credit card required. Paid plans bill per completed task. You also need your own Pinecone account.

Why does Pinecone need an index host?

Pinecone sends data requests to the host of a specific index, not to one shared address. You copy that host from the index page in the Pinecone console and paste it next to your API key. The AI then works with that index and its namespaces.