# Connect pi

https://omnimodel.me/apps/pi

> Minimal open-source terminal coding agent; custom endpoints live in models.json.

## Use cases

For anyone who wants a small, extensible terminal coding agent. pi keeps custom endpoints in `~/.pi/agent/models.json`; `api: openai-completions` selects the chat/completions protocol.

## Tips

A coding agent sends many long requests per task, so pick a long-context model with tool calling and set a daily spend limit on the key.

## Reasoning effort

Whether a reasoning level is sent depends on the reasoning setting you give the model (`reasoning_effort` on the Chat API). The platform follows the levels the serving line accepts: a level the upstream rejects becomes the next higher accepted level, or its highest one. If the model cannot turn reasoning off, asking for it returns 422 `model.reasoning_off_unsupported`. Each model's table is under Reasoning effort on its page in the [catalog](/catalog?lang=en); rewritten calls show the rewrite in usage records.

## If it cannot connect

1. Set `baseUrl` to `https://omnimodel.me/v1`, without `/chat/completions`.
2. The key starts with `key_` and is not paused or revoked; `auth.invalid_key` means the key cannot be used.
3. The model ID matches the [catalog](/catalog) and is in the key's allowed models.
4. `billing.insufficient_credit` means the balance is too low; see [common errors](/docs/errors?lang=en) for more.

## Setup

1. Install: npm install -g --ignore-scripts @earendil-works/pi-coding-agent (Node.js 22.19 or later).
2. Put the block below in ~/.pi/agent/models.json.
3. Run pi --model a2agent/{{model}} in your project, or switch with /model inside a session.

## Settings

- API host：`https://omnimodel.me/v1`
- API key：`YOUR_API_KEY`
- Model：`MODEL_ID`

> This setup follows the pi docs and has not been tested by the platform yet; if something breaks, open a ticket with the request ID.

```json
{
  "providers": {
    "a2agent": {
      "name": "A2Agent",
      "baseUrl": "https://omnimodel.me/v1",
      "api": "openai-completions",
      "apiKey": "YOUR_API_KEY",
      "models": [{ "id": "MODEL_ID" }]
    }
  }
}
```

## FAQ

### The context size looks wrong. What now?

pi assumes 128K context and 16K output for custom models. If yours differs, add contextWindow and maxTokens to the model entry.

### Do I need to restart after editing models.json?

No. Opening /model inside a session reloads it.