Provider REST API

10 actions

Connect Claude to OpenAI

Call models, embeddings, files, and vector stores. Toolspoke puts 10 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: 10
Cost per call (typical): 1 credit
Adapter: Maintained by Toolspoke

Connected in three steps

1
Install OpenAI
Open the marketplace in your workspace, add OpenAI 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. OpenAI 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 OpenAI 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: platform.openai.com → API keys; sk-proj-…
Organization IDOptional: Only needed if the key belongs to several organizations; org-…
Project IDOptional: Scopes usage to one project; proj_…

What Claude can do in OpenAI

10 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: 8Reads
Writes: 2Writes
Destructive: 0Destructive

Reads

Fetches data and changes nothing.

list_models
List the models this API key can call. Use before `responses` or `embeddings` when you are unsure a model id exists for this account. Optionally filter by substring; returns at most `limit` entries.
embeddings
Create embedding vectors for one or more strings, for similarity search or clustering. Defaults to 256 dimensions to keep the response readable - raise `dimensions` when you are storing the vectors.
moderate
Classify text (and image URLs) against OpenAI's safety categories before you publish or forward it. Returns per-input `flagged`, category booleans, and confidence scores.
list_files
List files uploaded to this OpenAI account, newest first. Use it to find the file id for retrieve_file or to check what a vector store was built from.
retrieve_file
Fetch one file's metadata: name, bytes, purpose, and status. Call download_file to read its contents.
download_file
Download a file's contents as text, truncated to 10000 characters. Only useful for text formats such as .txt, .md, .json, or .jsonl - call retrieve_file first for its name, size, and purpose.
list_vector_stores
List the vector stores on this account with their file counts and sizes. Use it to find the vector_store_id that search_vector_store needs.
search_vector_store
Semantic search over the files in one vector store. Returns the matching chunks with their scores and source filenames - use this to ground an answer in the account's own documents.

Writes

Creates or updates something on the other side.

responses
Generate a model response with the Responses API. Use for text generation, structured JSON output, reasoning, and built-in tools like web_search. Returns the full response object plus a flattened output_text field. Streaming is not supported here.
generate_image
Generate images from a text prompt. The default dall-e-3 returns a temporary URL, which is cheap to pass around; gpt-image-* models return base64 bytes, which are omitted unless you set return_base64.

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 10 actions above are the whole of it. A call to any other name is refused before it reaches OpenAI 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.
Nothing here deletes: This connector writes to OpenAI, but nothing in it deletes or permanently alters anything.
It reaches no further than your credential: Toolspoke holds no access to OpenAI 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 OpenAI: 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 OpenAI. An agent has to ask.
It does not smooth over provider limits: Toolspoke does not retry, queue or back off around OpenAI's own rate limits. A call that OpenAI refuses comes back to the agent as a failed call.

Before you connect it

What can Claude do in OpenAI?

10 named actions: 8 that only read and 2 that write. They include responses, list_models and embeddings. 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 OpenAI connector need?

API key. The connector asks for api key, and optionally organization id and project id. 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 OpenAI connector work with Cursor and Codex, or only Claude?

Any client that speaks MCP, and every one of them gets the same 10 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 OpenAI connector not do?

The 10 actions above are the whole of it. A call to any other name is refused before it reaches OpenAI 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. This connector writes to OpenAI, but nothing in it deletes or permanently alters anything. Toolspoke holds no access to OpenAI 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 OpenAI. An agent has to ask. Toolspoke does not retry, queue or back off around OpenAI's own rate limits. A call that OpenAI 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 OpenAI'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 OpenAI?

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 the credits spent. 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 ai connectors

Browse all of them