Skip to content
actionlake
Start free

Connectors · AI

Pinecone MCP server for AI agents.

actionlake is an MCP gateway: one remote MCP server that connects Claude and other MCP clients to Pinecone and the rest of your systems. It gives your agent 52 Pinecone actions (27 reads, 17 writes and 8 that wait for a confirmation) and holds the Pinecone credential itself, so the agent never does. It paces Pinecone's limits and records every call.

Connect Pinecone freeSee pricing

100,000 free calls a month. No card, no email, no password: press Start a free sandbox on the next screen and keep the key it shows you.

claude mcp add --transport http actionlake https://connect.actionlake.com/mcp
the first call
do {  action: "pinecone.indexes.list",  intent: "check which Pinecone account is connected"}# List indexes · read · 1 upstream call
  • 52actionsin one signed file
  • 27readlook, never change
  • 17writecreate and update
  • 8wait for a yesdestructive, confirmed first

Set up

Connect Pinecone in three steps.

  1. 01

    Add actionlake to Claude.

    Run the command once in Claude Code; its consent page can create a free sandbox: 100,000 calls a month, no card. In the Claude app, add a custom connector with the URL below and paste your key on the consent page. Creating a new sandbox from the Claude app is not yet tested by us.

    https://connect.actionlake.com/mcp
  2. 02

    Connect Pinecone.

    API key. No consent screen: your agent can connect it for you, but the key then passes through your chat. To keep it out of the model's context, paste it in the console instead. actionlake stores what you paste write-only, sealed under your tenant's key.

    • api_keyPinecone API key: console (app.pinecone.io) > your project > API keys > Create API key
    • hostWhere your Pinecone lives: your host
    account { op: "connections.requirements", service: "pinecone" }
# free: exactly what to paste, and what is found for you
  3. 03

    Make the first call. Check before a write.

    Ask in plain words; Claude picks the action. Before anything changes, check validates the call and prices it, for free.

    do    { action: "pinecone.indexes.list" }# list indexes · read · 1 upstream callcheck { action: "pinecone.indexes.create", input: { … } }# free: inputs, permissions, cost and confirmations, nothing sent

Pinecone at a glance

Sign-in
API key. No consent screen: your agent can connect it for you, but the key then passes through your chat. To keep it out of the model's context, paste it in the console instead.
You provide
api_key
Setup fields
host
Limits we pace to
20 calls a second, per connection
API covered
Pinecone API 2025-10 (db_control, db_data, inference, assistant_control, assistant_data, assistant_evaluation), at your Pinecone instance's API base URL, which you enter when connecting
Built from
Pinecone's documentation, read 22 October 2026OpenAPI spec (GitHub)OpenAPI spec (GitHub)OpenAPI spec (GitHub)
Service file
Version 1.0.0, signed (Ed25519)
Cached reads
27 of 27 reads, for up to 2 minutes; a write clears what it affects
Shared entities
1, so ask can line Pinecone up with your other systems
Data
Processed and stored in Sydney, Australia. No US region yet

Capabilities

What an AI agent can do in Pinecone.

52 actions in 14 areas. Every one carries its effect, so a key can allow the reads and hold the rest.

  • 27readlook, never change
  • 17writecreate and update
  • 8destructivewait for a confirmation

  1. Assistants

    8

    3 read · 4 write · 1 destructive

    Chat with an assistant · Chat through an OpenAI-compatible interface · Retrieve context from an assistant · Create an assistant

  2. Vectors

    7

    4 read · 2 write · 1 destructive

    Fetch vectors by metadata · List vector IDs · Search with a vector · Fetch vectors

  3. Backups

    6

    3 read · 2 write · 1 destructive

    Create a backup of an index · Create an index from a backup · List backups for all indexes in a project · List backups for an index

  4. Indexes

    6

    2 read · 3 write · 1 destructive

    List indexes · Create an index · Create an index with integrated embedding · Describe an index

  5. Bulk imports

    4

    2 read · 1 write · 1 destructive

    List imports · Start import · Describe an import · Cancel an import

  6. Collections

    4

    2 read · 1 write · 1 destructive

    Create a collection · List collections · Describe a collection · Delete a collection

  7. Namespaces

    4

    2 read · 1 write · 1 destructive

    Create a namespace · List namespaces · Describe a namespace · Delete a namespace

  8. Assistant files

    3

    2 read · 0 write · 1 destructive

    List Files · Describe a file upload · Delete an uploaded file

6 more areas · 10 actions
  • Inference 2
  • Models 2
  • Records 2
  • Restore jobs 2
  • Evaluation 1
  • Index stats 1

Top actions

The Pinecone API, shaped for agents.

The ids your agent types, what each does, and the one name it shares with other systems.

ActionWhat it doesEffectSame name in
pinecone.indexes.listList indexesList indexes.readthis connector
pinecone.vectors.listList vector IDsList vector IDs.readthis connector
pinecone.vectors.querySearch with a vectorSearch with a vector.readthis connector
pinecone.indexes.createCreate an indexCreate an index.writethis connector
pinecone.records.searchSearch with textSearch with text.readthis connector
pinecone.assistants.chatChat with an assistantChat with an assistant.writethis connector
pinecone.assistants.listList assistantsList assistants.readthis connector
pinecone.inference.embedGenerate vectorsGenerate vectors.writethis connector
pinecone.namespaces.listList namespacesList namespaces.readthis connector
pinecone.inference.rerankRerank resultsRerank results.readthis connector
pinecone.namespaces.createCreate a namespaceCreate a namespace.writethis connector
pinecone.indexes.deleteDelete an indexDelete an index.destructivethis connector

Try it

Prompts to paste into Claude.

Each one names the Pinecone actions it will call, so you know what it reads and what it changes before you send it.

  1. Readread

    List indexes in Pinecone and tell me what stands out.

    callspinecone.indexes.list

  2. Writewrite

    Create an index in Pinecone. Show me exactly what you will send before you save it.

    callspinecone.indexes.create

Governance for Pinecone

Destructive actions wait for a yes.

8 Pinecone actions are destructive. These are all of them:

  • Delete an indexpinecone.indexes.delete
  • Delete vectorspinecone.vectors.delete
  • Delete an assistantpinecone.assistants.delete
  • Delete a namespacepinecone.namespaces.delete
  • Delete a backuppinecone.backups.delete
  • Delete a collectionpinecone.collections.delete
  • Delete an uploaded filepinecone.assistant_files.delete
  • Cancel an importpinecone.bulk_imports.cancel

A destructive action runs only with confirm: true. Under the human policy the agent cannot confirm for itself: the run stops until a named person approves, in the console or from a signed link in an email. The cap on a key is refused before the call is sent, never after.

dopinecone.indexes.deleteconfirm_policy: human

Delete an index?

intent: the user asked me to delete an index

run suspended · waiting for a named person · key pinecone-agent cannot confirm itself
In the console, or from an emailed signed link.
a key for this connector
account {  op: "keys.create", name: "pinecone-agent",  allowed_actions: ["pinecone.*"], cap_calls: 2000,  confirm_policy: "human_for_writes",  approval_rules: [    { match: "pinecone.indexes.delete", approver: "[email protected]" },    { match: "pinecone.vectors.delete", approver: "[email protected]" },    { match: "pinecone.assistants.delete", approver: "[email protected]" }  ]}
Illustrative. Scoped to Pinecone only, capped at 2,000 upstream calls, every write confirmed, and the destructive ones approved by a named person.

Cost

What Pinecone costs on actionlake.

Every Pinecone action is one upstream call, the only thing actionlake bills. Searching, check, cost, account and any call Pinecone refused with 401, 403, 429 or a 5xx are free. Paid plans start at US$49 a month for 250,000 calls, with overage from US$0.30 down to US$0.18 per 1,000.

  • 1upstream call per actionall 52 Pinecone actions
  • 6actions pageeach page of a long list is one call
  • 27reads can be cacheda cache hit costs nothing
  • 100,000calls a month, freepaid plans from US$49 a month for 250,000

See pricing

Action index

Every Pinecone action.

52 actions by area: the id your agent calls, what it does, and its effect.

Assistants8
  • pinecone.assistants.chatChat with an assistantwrite
  • pinecone.assistants.chat_completionsChat through an OpenAI-compatible interfacewrite
  • pinecone.assistants.contextRetrieve context from an assistantread
  • pinecone.assistants.createCreate an assistantwrite
  • pinecone.assistants.deleteDelete an assistantdestructive
  • pinecone.assistants.getCheck assistant statusread
  • pinecone.assistants.listList assistantsread
  • pinecone.assistants.updateUpdate an assistantwrite
Vectors7
  • pinecone.vectors.deleteDelete vectorsdestructive
  • pinecone.vectors.fetchFetch vectorsread
  • pinecone.vectors.fetch_by_metadataFetch vectors by metadataread
  • pinecone.vectors.listList vector IDsread
  • pinecone.vectors.querySearch with a vectorread
  • pinecone.vectors.updateUpdate a vectorwrite
  • pinecone.vectors.upsertUpsert vectorswrite
Backups6
  • pinecone.backups.createCreate a backup of an indexwrite
  • pinecone.backups.create_indexCreate an index from a backupwrite
  • pinecone.backups.deleteDelete a backupdestructive
  • pinecone.backups.getDescribe a backupread
  • pinecone.backups.listList backups for all indexes in a projectread
  • pinecone.backups.list_for_indexList backups for an indexread
Indexes6
  • pinecone.indexes.configureConfigure an indexwrite
  • pinecone.indexes.createCreate an indexwrite
  • pinecone.indexes.create_for_modelCreate an index with integrated embeddingwrite
  • pinecone.indexes.deleteDelete an indexdestructive
  • pinecone.indexes.getDescribe an indexread
  • pinecone.indexes.listList indexesread
Bulk imports4
  • pinecone.bulk_imports.cancelCancel an importdestructive
  • pinecone.bulk_imports.getDescribe an importread
  • pinecone.bulk_imports.listList importsread
  • pinecone.bulk_imports.startStart importwrite
Collections4
  • pinecone.collections.createCreate a collectionwrite
  • pinecone.collections.deleteDelete a collectiondestructive
  • pinecone.collections.getDescribe a collectionread
  • pinecone.collections.listList collectionsread
Namespaces4
  • pinecone.namespaces.createCreate a namespacewrite
  • pinecone.namespaces.deleteDelete a namespacedestructive
  • pinecone.namespaces.getDescribe a namespaceread
  • pinecone.namespaces.listList namespacesread
Assistant files3
  • pinecone.assistant_files.deleteDelete an uploaded filedestructive
  • pinecone.assistant_files.getDescribe a file uploadread
  • pinecone.assistant_files.listList Filesread
Inference2
  • pinecone.inference.embedGenerate vectorswrite
  • pinecone.inference.rerankRerank resultsread
Models2
  • pinecone.models.getDescribe a modelread
  • pinecone.models.listList available modelsread
Records2
  • pinecone.records.searchSearch with textread
  • pinecone.records.upsertUpsert textwrite
Restore jobs2
  • pinecone.restore_jobs.getDescribe a restore jobread
  • pinecone.restore_jobs.listList restore jobsread
Evaluation1
  • pinecone.evaluation.alignmentEvaluate an answerwrite
Index stats1
  • pinecone.index_stats.getGet index statsread

Questions

The Pinecone MCP server, answered.

Is there a Pinecone MCP server for Claude?

Yes. actionlake is an MCP gateway: a remote MCP server that gives Claude, or another MCP client, 52 Pinecone actions across assistants, vectors, backups and indexes, through the same eight tools it uses for every other connector. actionlake is independent; it is not made or endorsed by Pinecone.

What do I need to connect Pinecone?

Pinecone API key: console (app.pinecone.io) > your project > API keys > Create API key. Plus your host. The free call account { op: "connections.requirements", service: "pinecone" } lists exactly what to paste.

Can an AI agent delete or change things in Pinecone?

Only if its key allows it, and only with an explicit confirmation. 8 Pinecone actions are destructive, including delete an index, delete vectors and delete an assistant. Each runs only with confirm: true. Under the human confirm policy, or an approval rule on the key, the run waits until a named person approves it.

Will an AI agent use up my Pinecone API limit?

actionlake paces every call to the limits in the Pinecone service file: 20 calls a second per connection. Repeated reads are answered from cache (27 Pinecone actions can be), which never reaches Pinecone.

What does a Pinecone action cost?

One upstream call, the only unit actionlake bills. A list that pages bills one call per page (6 Pinecone actions page). Cache hits, searches, dry runs and calls Pinecone refused with 401, 403, 429 or a 5xx are free. The Free plan includes 100,000 calls a month with no card. Paid plans start at US$49 a month for 250,000 calls, with overage from US$0.30 down to US$0.18 per 1,000.

Which Pinecone API does it cover?

The Pinecone API 2025-10 (db_control, db_data, inference, assistant_control, assistant_data, assistant_evaluation) at your Pinecone instance's API base URL, which you enter when connecting. It is built from Pinecone's documentation as read on 22 October 2026. 52 actions in 14 areas, each one in the signed service file (version 1.0.0).

Does actionlake keep my Pinecone data, and where?

actionlake processes and stores data in Sydney, Australia; there is no US region yet. Reads may be cached for up to 2 minutes, and a write clears the reads it affects. Each run's request and response is kept, sealed, for your plan's trace window: 1 day on Free, 3 days on Starter, 7 days on Pro, 14 days on Team, 30 days on Business and 30 days on Enterprise. Your Pinecone credential is sealed under your tenant's key, never logged and never returned. Your traffic is not used to train models.

Which AI clients can use it?

Claude Code, tested by us end to end, and the Claude app as a custom connector, tested by us with an existing key. Other MCP clients that support remote servers over Streamable HTTP (MCP revisions 2025-06-18 and 2025-11-25) and OAuth dynamic client registration, or a bearer header, should connect the same way; we have not tested them. Or plain REST, under the same action names.

More connectors like Pinecone

Joins with, on shared entities

Browse

actionlake is independent. It is not made or endorsed by Pinecone; the name is the vendor's own. Counts are from the Pinecone service file, version 1.0.0.

Give your agent Pinecone today.

Connect Pinecone freeSee pricing

Start free - 100,000 calls/month, no card. No signup. Hard caps on every key. Cached answers are free.