Provider REST API

18 actions

Connect Claude to Pinecone

Manage Pinecone indexes, query and write vectors and records, and embed or rerank text with Pinecone's hosted models. Toolspoke puts 18 of its actions behind one MCP endpoint that Claude, Cursor, and Codex all speak.

Connect View pricing

Connection: Provider REST API
Authentication: API key
Actions exposed: 18
What a call bills: Nothing
Adapter: Maintained by Toolspoke

Connected in three steps

1
Install Pinecone
Open the marketplace in your workspace, add Pinecone to the project your agents work in, and it appears on the gateway immediately.
2
Connect the credential
Authenticate with api key. Where to get one, and what it has to be able to reach, is the next section.
3
Point your agent at the gateway
Give your client one address, https://toolspoke.com/mcp. Claude Code takes it as a command, Claude and Claude Desktop add it as a custom connector, and Cursor, Codex and VS Code each read it from a config file of their own.

.mcp.json

{
  "mcpServers": {
    "toolspoke": {
      "type": "http",
      "url": "https://toolspoke.com/mcp"
    }
  }
}

One block covers every tool you have installed. Pinecone shows up in the client as soon as your policy allows it, and so does everything else you install later.

Where the address goes, per client

Claude Code: Run it in your project, then /mcp to sign in
claude mcp add --transport http toolspoke https://toolspoke.com/mcp
Claude and Claude Desktop: Settings, then Connectors, then Add custom connector
https://toolspoke.com/mcp
Cursor: ~/.cursor/mcp.json, or .cursor/mcp.json for one project
{ "mcpServers": { "toolspoke": { "url": "https://toolspoke.com/mcp" } } }
Codex: ~/.codex/config.toml
[mcp_servers.toolspoke] url = "https://toolspoke.com/mcp"
VS Code: .vscode/mcp.json, or the MCP: Add Server command
{ "servers": { "toolspoke": { "type": "http", "url": "https://toolspoke.com/mcp" } } }

What Pinecone asks for

API key. You provide it once, when you install the connector. Toolspoke encrypts it at rest and decrypts it only for the length of a single call, and the gateway attaches it to the outbound request itself, so it is never part of the arguments an agent sends.

API keyRequired: Pinecone console → choose the project → API keys → Create API key. A key reaches only the project that issued it. Give it All permissions for writes, or ReadOnly to use only the reads here.; pcsk_…

What Claude can do in Pinecone

18 actions, each one declared and named by the connector rather than discovered at runtime. A workspace policy grants a person all of them, a hand-picked selection, everything on the read side, everything on the write side, or none.

Reads: 11Reads
Writes: 4Writes
Destructive: 3Destructive

Reads

Fetches data and changes nothing.

list_indexes
List every index in the project with its dimension, metric, embedding model (for integrated indexes) and `host`. The host is what every vector and record action takes as index_host.
describe_index
Read one index's configuration, status and host. For an index with integrated embedding, `embed.field_map` names the record field that holds the text.
list_models
List the embedding and reranking models Pinecone hosts, with their dimensions and limits.
describe_index_stats
Count the vectors in an index, per namespace, and report its dimension and fullness. Use it to find which namespaces exist.
list_namespaces
List the namespaces in an index with their record counts.
query_vectors
Find the nearest vectors to a query vector, or to a stored vector by id. The vector must come from the same model the index was built with; for integrated indexes, search_records takes text instead.
search_records
Search an index with integrated embedding by text. Pinecone embeds the query with the index's own model, and can rerank the hits.
fetch_vectors
Read vectors by id, with their values and metadata.
list_vector_ids
List vector ids in a namespace, optionally only those starting with a prefix. Serverless indexes only.
embed
Turn text into vectors with one of Pinecone's hosted embedding models. Stores nothing.
rerank
Order documents by relevance to a query with a hosted reranking model. Stores nothing.

Writes

Creates or updates something on the other side.

create_index_for_model
Create a serverless index with integrated embedding, so records are embedded by Pinecone on upsert and searched by text. `embed.field_map.text` names the record field that holds the text.
configure_index
Change an index's deletion protection or tags.
upsert_vectors
Write vectors, replacing any with the same id. Each needs an id and values of the index's dimension; metadata is optional.
update_vector
Change one vector's values, or set metadata keys on it. Keys not named are left as they are.

Destructive

Deletes or permanently alters something. Worth granting on purpose.

delete_vectors
Delete vectors by id, by metadata filter, or every vector in a namespace. There is no undo.
delete_namespace
Delete a namespace and every record in it. There is no undo. Serverless indexes only.
delete_index
Delete an index and all its data. There is no undo, and Pinecone refuses while deletion protection is on.

What it will not do

Enforced by the gateway rather than left to convention, which is why each of these can be stated flatly.

It cannot call anything else: The 18 actions above are the whole of it. A call to any other name is refused before it reaches Pinecone rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time.
It reaches no further than your credential: Toolspoke holds no access to Pinecone of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either.
It never hears from Pinecone: Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Pinecone. An agent has to ask.
It does not smooth over provider limits: Toolspoke does not retry, queue or back off around Pinecone's own rate limits. A call that Pinecone refuses comes back to the agent as a failed call.

Before you connect it

What can Claude do in Pinecone?

18 named actions: 11 that only read, 4 that write and 3 that delete or permanently alter something. They include list_indexes, describe_index and list_models. Nothing outside that list is reachable: the connector declares each operation by name rather than proxying whatever an agent asks for.

What credentials does the Pinecone connector need?

API key. The connector asks for api key. Values are encrypted at rest and attached to the outbound request by the gateway, so they are never part of the arguments an agent sends and never reach the audit log.

Does the Pinecone connector work with Cursor and Codex, or only Claude?

Any client that speaks MCP, and every one of them gets the same 18 actions. There is a single address, https://toolspoke.com/mcp. Claude Code adds it with claude mcp add --transport http, Claude and Claude Desktop take it as a custom connector in settings, Cursor reads it from .cursor/mcp.json, Codex from ~/.codex/config.toml, and VS Code from .vscode/mcp.json. Each of them signs in to the gateway itself, so there is no key to paste.

What does the Pinecone connector not do?

The 18 actions above are the whole of it. A call to any other name is refused before it reaches Pinecone rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time. Toolspoke holds no access to Pinecone of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either. Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Pinecone. An agent has to ask. Toolspoke does not retry, queue or back off around Pinecone's own rate limits. A call that Pinecone refuses comes back to the agent as a failed call.

Can I limit which actions an agent can call?

Yes, in two places. The project switches Pinecone's actions on and off one at a time, for everyone in the project at once, and the screen groups them by read, write and destructive so turning off everything that deletes is one click. An individual agent key can then be narrowed further, to particular toolkits in a project and to particular actions in a toolkit. Whatever it was granted, a key never reaches a project its owner cannot.

What gets recorded when an agent calls Pinecone?

Every attempt, with the agent that made it and the person that agent belongs to, the full request payload, the response payload, the status, the duration, and what the call cost in sandbox time and model tokens. Values whose key names a secret are masked out before the record is shown to anyone. An operation the connector marks as not retained never has its response body written at all, so the gateway keeps no second copy of what was read.

More database connectors

Browse all of them