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

# Embed with your coding agent

> Install the opengeni-client skill and let Claude Code, Codex, or Cursor add an OpenGeni agent to your product.

The `opengeni-client` skill teaches a coding agent how to integrate OpenGeni the right way: the full conversation behind a tenant-scoped backend proxy, explicit user onboarding, a minimal tool surface, and your own tools for data. Give it to your agent and describe what you want.

## What you need

* An OpenGeni organization. Sign up at [app.opengeni.ai](https://app.opengeni.ai) or [self-host](/guides/self-host).
* A **full-access organization API key**: **Organization settings → Developer → API keys**. The token is shown once; store it in your secret manager.
* Three environment variables on your backend:

| Variable | Value |
| - | - |
| `OPENGENI_API_BASE_URL` | `https://app.opengeni.ai`, or your self-hosted API URL |
| `OPENGENI_ORGANIZATION_ID` | **Organization settings → General → Organization ID** |
| `OPENGENI_API_KEY` | The organization API key. Server only; never send it to the browser |

## Copy this prompt

Paste it into Claude Code, Codex, Cursor, or any agent that can run commands in your repository:

```text theme={null}
Add an OpenGeni agent to this product.

1. Install the OpenGeni integration skill:
     npx skills add Cloudgeni-ai/opengeni --skill opengeni-client
   If that is unavailable, copy
   https://github.com/Cloudgeni-ai/opengeni/tree/main/.agents/skills/opengeni-client
   into this repository's skills directory, or read the full text at
   https://docs.opengeni.ai/reference/opengeni-client-skill.md
2. Read the skill, then inspect this repository: auth, tenants, data routes,
   frontend, tests.
3. Embed the default OpenGeni conversation (OpenGeniChat, or
   SessionConversation for one record, behind createSessionProxyHandler with
   the framework adapter) where it fits the product, and give the agent our
   data through our own authenticated tools.
4. Before building, ask me once about anything the repository does not
   answer: who shares chats, what runs on a schedule, where results land,
   and whether the agent may write.

Credentials are in the environment: OPENGENI_API_BASE_URL,
OPENGENI_ORGANIZATION_ID, OPENGENI_API_KEY. Keep the API key on the server.
```

You can also install the skill yourself first:

<CodeGroup>
  ```bash npx theme={null}
  npx skills add Cloudgeni-ai/opengeni --skill opengeni-client
  ```

  ```bash bunx theme={null}
  bunx skills add Cloudgeni-ai/opengeni --skill opengeni-client
  ```
</CodeGroup>

## What you get

A typical integration adds:

* **Onboarding.** Each of your tenants maps to an OpenGeni workspace, and each admitted user becomes a member with only the permissions the conversation needs.
* **One backend route.** `createSessionProxyHandler` mounted at `/api/opengeni/*` through the Next.js, Express, or Hono adapter, authenticating every request with your existing session check.
* **The conversation.** `OpenGeniProvider` and `OpenGeniChat` (the user's chats plus the conversation), or `SessionConversation` for one record, styled with your brand tokens.
* **Your tools.** An MCP server or OpenAPI Integration that exposes the product data and actions the agent may use, authorized as the signed-in user.
* **Tests** for the proxy and for tenant isolation.

The agent asks before choosing things only you can decide, such as whether chats are private or shared and whether the agent may change data.

## Next steps

<CardGroup cols={2}>
  <Card title="Embed manually" icon="code" href="/embed-manually">
    The same integration in four steps, if you prefer to write it yourself.
  </Card>

  <Card title="How it fits together" icon="diagram-project" href="/integrate/how-it-fits-together">
    What the agent builds, and why each piece is there.
  </Card>
</CardGroup>