> ## Documentation Index
> Fetch the complete documentation index at: https://docs.askalchemist.com/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenAI Responses API

> Add Alchemist as a hosted MCP tool on the OpenAI Responses API.

The OpenAI Responses API can call Alchemist server-side via its hosted `mcp` tool — OpenAI's infrastructure connects to the endpoint on the model's behalf, so there's nothing to run in your own process. Pass your API key in the `authorization` field.

Get a key at [app.askalchemist.com/settings/api-keys](https://app.askalchemist.com/settings/api-keys).

```json theme={null}
{
  "model": "<your-model>",
  "tools": [
    {
      "type": "mcp",
      "server_label": "alchemist",
      "server_url": "https://api.askalchemist.com/mcp/",
      "authorization": "alch_YOUR_KEY",
      "require_approval": "never"
    }
  ],
  "input": "Did the March 2026 FOMC minutes signal a rate pause?"
}
```

OpenAI forwards `authorization` as the `Authorization` header and does not store it — send it on each request. Optional fields: `allowed_tools` (restrict which tools are exposed) and `require_approval` (`"never"` | `"always"` | per-tool object).

<Note>
  The Responses API expects a token you already hold — it doesn't run the OAuth 2.0 + Dynamic Client Registration flow for you. The static API key is the clean fit here.
</Note>