# Build a spinach.ai agent with Claude Code

> Have Claude Code build a spinach.ai agent with the Agent200 SDK: managed OAuth, ready-made tools and one API key.

- Page: https://www.agent200.com/connectors/spinach-ai/framework/claude-code/
- Connector: [spinach.ai](https://www.agent200.com/connectors/spinach-ai.md)
- Framework: Claude Code
- Updated: 2026-11-18

## Instructions for AI agents

Goal: connect spinach.ai to Claude Code with Agent200.
Guide: https://www.agent200.com/connectors/spinach-ai/framework/claude-code/

1. Get an Agent200 API key. If the developer has none, send them to https://agent200.com/platform/apikeys/ and ask them to choose "Create an API key". Read the key from the AGENT200_API_KEY environment variable. Never ask for it in chat, and never write it into code, prompts or logs.
2. Make sure spinach.ai is enabled for the project in the Agent200 dashboard (Connectors), with the endpoints and scopes the task needs.
3. Add the Agent200 SDK to the project and create a client with that key.
4. Give every request a stable id for the end user. Agent200 uses only that user's own connections.
5. Connect the user's spinach.ai account through Agent200. If Agent200 returns an authorization link, show it to the user and continue once they have approved access on spinach.ai.
6. Run requests through Agent200 with a model (for example openai/gpt-5.6-sol) and "spinach_ai" in the tools. The model decides when a spinach.ai tool is needed; Agent200 executes the call.
7. Check it works with this request: "What were the action items from this morning's standup"
   What should come back:
   - A result from that user's own spinach.ai account.
   - Behind the scenes, Claude Code calls `spinach_ai_list_meetings`.
8. If a tool is missing or a call is refused, check the project's Connectors settings in the dashboard and the access the user granted in spinach.ai, then retry.

The code on the guide page is a conceptual example of the SDK. SDK reference: https://docs.igpt.ai/

## Overview

This guide has Claude Code build a spinach.ai agent with the Agent200 SDK, in your own project. When you finish, your agent can call the spinach.ai tools for each of your users.

Claude Code writes the code from the steps below; Agent200 runs the spinach.ai calls with each user's own connection and handles the OAuth flow and the tokens. You do not build or maintain a spinach.ai integration.

- **Connector:** [spinach.ai](https://www.agent200.com/connectors/spinach-ai/)
- **Works with:** Claude Code · Coding agent, with the SDK
- **Authentication:** Managed OAuth, per end user
- **You need:** [An Agent200 API key](https://agent200.com/platform/apikeys/)
- **Tools:** [4 tools](https://www.agent200.com/connectors/spinach-ai/framework/claude-code/#tools)
- **Model in examples:** `openai/gpt-5.6-sol`

## Before you start

- **An Agent200 API key.** Create one on the [API keys page](https://agent200.com/platform/apikeys/) of the dashboard.
- **spinach.ai enabled for your project.** In the dashboard's Connectors settings, with the endpoints and scopes your agent needs.
- **An end user with a spinach.ai account.** They approve access once, on spinach.ai's own page.
- **Claude Code in your project.** It writes the code with the Agent200 SDK; you review it.

## Connect spinach.ai to Claude Code

### Step 1: Create your Agent200 API key

Open the API keys page in the Agent200 dashboard and choose Create an API key. Copy the key and store it as an environment variable, AGENT200_API_KEY, or in your secrets manager. One key covers spinach.ai and every other connector and model enabled for your project. Keep it out of chats, prompts and source control.

Create an API key: https://agent200.com/platform/apikeys/

File `.env`:

```bash
# Keep the key out of your code, prompts and chats.
export AGENT200_API_KEY="paste-your-key-here"
```

(Conceptual example of the Agent200 SDK.)

### Step 2: Give this guide to Claude Code

Start Claude Code in your project and paste the agent brief from this page, or give it the address of this guide's Markdown version. Claude Code follows the steps below in order: the SDK, the spinach.ai connection and the first request. Review the code it writes before you run it.

### Step 3: Set up the Agent200 SDK

Add the SDK to your project and create a client with the key from the environment.

File `agent.js`:

```js
// One Agent200 API key covers every connector enabled for the project.
const agent200 = new Agent200({
  apiKey: process.env.AGENT200_API_KEY,
});
```

(Conceptual example of the Agent200 SDK.)

### Step 4: Let your user connect spinach.ai

Each end user authorizes their own spinach.ai account once, through managed OAuth. If the user has not connected yet, Agent200 returns an authorization link: show it to them, and they approve access on spinach.ai's own page. Agent200 handles the callback, stores the tokens and refreshes them.

File `connect.js`:

```js
// Each end user authorizes their own spinach.ai account, once.
const connection = await agent200
  .user("user_123")
  .connect("spinach_ai");

if (connection.authorizationUrl) {
  // Show this link to your user; Agent200 handles the rest.
  console.log(connection.authorizationUrl);
}
```

(Conceptual example of the Agent200 SDK.)

### Step 5: Run with the spinach.ai tools

Name the user, the model and spinach.ai in the tools. The model decides when a spinach.ai tool is needed, and Agent200 executes the call with the user's own connection.

File `run.js`:

```js
const result = await agent200
  .user("user_123")
  .run({
    model: "openai/gpt-5.6-sol",
    input: "What were the action items from this morning's standup",
    tools: ["spinach_ai"],
  });

console.log(result.output);
```

(Conceptual example of the Agent200 SDK.)

## Test your connection

Ask: "What were the action items from this morning's standup"

What should come back:

- A result from that user's own spinach.ai account.
- Behind the scenes, Claude Code calls `spinach_ai_list_meetings`.

If you get an authorization link instead, the user has not connected spinach.ai yet. Show them the link, then run the request again.

## spinach.ai tools Claude Code can call

- **List meetings** (`spinach_ai_list_meetings`, read): List the user's recorded meetings.
- **Get a summary** (`spinach_ai_get_summary`, read): Read the summary of a meeting.
- **Get action items** (`spinach_ai_get_action_items`, read): Read the action items from a meeting.
- **Get a transcript** (`spinach_ai_get_transcript`, read): Read the transcript of a meeting.

## How it works

1. **Your app sends the request.** It calls Agent200 with the model, instructions, the input, the end user's id and the spinach.ai tools.
2. **The model picks a tool.** The model decides whether a spinach.ai tool is needed, and with which arguments.
3. **Agent200 runs the call.** It uses that user's own spinach.ai connection, within the endpoints and scopes your project allows, and returns the result to the model.
4. **Your app gets the result.** Agent200 returns the completed result. Your app decides when the agent runs and what happens next.

**Each user's own connection.** Connections belong to individual end users. A request for one user never reaches another user's spinach.ai account.

**Permissions at every level.** Your project settings decide what the agent may request, the end user decides what they authorize, and spinach.ai enforces what its API grants.

## Troubleshooting

### The request says spinach.ai authorization is required

The end user has not connected spinach.ai yet, or the connection needs to be approved again. Agent200 returns an authorization link: show it to the user, let them approve access on spinach.ai's own page, then run the request again.

### Claude Code does not use the spinach.ai tools

Check that spinach.ai is enabled for your project in the dashboard's Connectors settings and that it is in the tools you send with the request. Naming spinach.ai in the request, or in the instructions, also helps the model choose the right tool.

### A tool call is refused or returns less than expected

Agent200 runs only what your project allows (services, endpoints and scopes), and spinach.ai returns only what the end user authorized. Widen the endpoints or scopes in the dashboard if the agent needs more, then ask the user to connect spinach.ai again so they can approve the new access.

### The results come from the wrong account

Every request names an end user (`user_123` in the examples). Pass the id of the person your agent is working for: Agent200 then uses only that user's spinach.ai connection, never another user's.

### The API key was pasted into a chat or committed

Create a new key on the [API keys page](https://agent200.com/platform/apikeys/), update `AGENT200_API_KEY` wherever your app reads it, and stop using the old one. Keep keys in environment variables or a secrets manager, never in prompts, chats or source control.

## Use spinach.ai with another framework

- [spinach.ai + Claude](https://www.agent200.com/connectors/spinach-ai/framework/claude.md)
- [spinach.ai + ChatGPT](https://www.agent200.com/connectors/spinach-ai/framework/chatgpt.md)
- [spinach.ai + Muse](https://www.agent200.com/connectors/spinach-ai/framework/muse.md)
- [spinach.ai + Grok Bot](https://www.agent200.com/connectors/spinach-ai/framework/grok-bot.md)
- [spinach.ai + OpenAI Dots](https://www.agent200.com/connectors/spinach-ai/framework/openai-dots.md)
- [spinach.ai + Claude Duet](https://www.agent200.com/connectors/spinach-ai/framework/claude-duet.md)
- [spinach.ai + Hermes Agent](https://www.agent200.com/connectors/spinach-ai/framework/hermes-agent.md)
- [spinach.ai + OpenClaw](https://www.agent200.com/connectors/spinach-ai/framework/openclaw.md)

---

Agent200: https://www.agent200.com/ · API keys: https://agent200.com/platform/apikeys/ · Docs: https://docs.igpt.ai/
