Pinecone for AI agents
Pinecone is a managed vector database used to store embeddings and power fast similarity search for AI applications. Developers use it to give AI features long-term memory and retrieval. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.
What agents do in Pinecone.
Query an index for similar vectors
Fetch the nearest matches to a query embedding from a namespace for a retrieval task.
Check index stats read-only
Look up an index's vector count and stats before deciding whether to expand capacity.
Delete vectors with approval
Remove a batch of vectors from a namespace only after confirming they are no longer needed.
48 Pinecone actions, graded by risk.
Every Pinecone action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
24Look things up. Allowed by default.
- pinecone.get_modelGet Model Information
- pinecone.list_filesList Files
- pinecone.list_modelsList Available Models
- pinecone.list_indexesList Indexes
- pinecone.list_vectorsList Vectors
- pinecone.get_assistantGet Assistant Status
- pinecone.query_vectorsQuery Vectors
- pinecone.list_assistantsList Assistants
- pinecone.list_collectionsList Collections
- pinecone.list_bulk_importsList Bulk Imports
- pinecone.list_restore_jobsList Restore Jobs
- pinecone.list_index_backupsList Index Backups
- pinecone.list_project_backupsList Project Backups
- pinecone.search_records_namespaceSearch Records in Namespace
- pinecone.list_namespaces_operationList Namespaces
- pinecone.describe_fileDescribe File
write
17Create and change things. Allow, or ask the user first.
- pinecone.create_indexCreate Index
- pinecone.create_backupCreate Backup
- pinecone.update_vectorUpdate Vector
- pinecone.create_assistantCreate Assistant
- pinecone.create_namespaceCreate Namespace
- pinecone.update_assistantUpdate Assistant
- pinecone.create_index_for_modelCreate Index with Embedding Model
- pinecone.create_index_from_backupCreate Index from Backup
- pinecone.embedGenerate Embeddings
- pinecone.rerankRerank Documents
- pinecone.upload_fileUpload File to Assistant
- pinecone.chat_assistantChat with Assistant
- pinecone.upsert_vectorsUpsert Vectors
- pinecone.configure_indexConfigure Index
- pinecone.start_bulk_importStart Bulk Import
- pinecone.upsert_records_namespaceUpsert Records to Namespace
destructive
7Delete, cancel or archive. Ask first, or deny outright.
- pinecone.delete_fileDelete File
- pinecone.delete_indexDelete Index
- pinecone.delete_backupDelete Backup
- pinecone.delete_vectorsDelete Vectors
- pinecone.delete_assistantDelete Assistant
- pinecone.delete_namespaceDelete Namespace
- pinecone.cancel_bulk_importCancel Bulk Import
Pinecone in three steps.
- 01Your users connect PineconeThey sign in to Pinecone on Arc0 Connect, under your brand, and approve the access you ask for.
- 02You set the rulesReads run, writes like “create Index” can wait for the user, and “delete File” can be denied outright.
- 03Any agent can actYour agent calls Pinecone through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('pinecone', { read: 'allow', write: 'ask', // create_index destructive: 'deny', // delete_file }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How Pinecone connects.
Users sign in to Pinecone on Arc0 Connect and approve the scopes you request. Build with Arc0’s Pinecone OAuth app, or bring your own so the Pinecone consent screen names you. Tokens refresh automatically, and you can export them whenever you want.
The same Pinecone connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →
- AUTH
- API key · OAuth 2.0
- CREDENTIALS
- Per-tenant encrypted vault
- MODEL SEES
- Results only, never credentials
- AUDIT LOG
- Every call, on every plan
Use Pinecone from any agent.
Pinecone and Arc0, answered.
Can I use Pinecone with Claude, ChatGPT or Cursor?
Yes. Connect Pinecone to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Pinecone actions you allow.
How do users connect Pinecone?
Users sign in to Pinecone on Arc0 Connect and approve the scopes you request. Build with Arc0’s Pinecone OAuth app, or bring your own so the Pinecone consent screen names you. Tokens refresh automatically, and you can export them whenever you want.
Which Pinecone actions can my agent take?
48 in total: 24 read, 17 write and 7 destructive, such as “create Index”. Your policies decide which of them each agent may call.
Can I stop my agent from deleting things in Pinecone?
Yes. Actions like “delete File” are graded destructive. Set destructive actions to deny, or to ask so the user approves each one, and blocked calls still show up on the audit log.
Can my own backend call Pinecone too?
Yes. The same Pinecone connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug Pinecone into your agent.
Your users connect Pinecone once, under your brand. Your agent gets 48 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan