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

# General Access

> Connect any compatible tool to ruoli with one recipe

ruoli is compatible with the API protocols of mainstream large models. As long as your tool lets you set a custom API URL and Key, it works — including tools not listed individually in these docs.

You only need two things:

| Item    | Value                                                                                   |
| ------- | --------------------------------------------------------------------------------------- |
| API Key | Create and copy it on the **Tokens** page in the [ruoli.dev](https://ruoli.dev) console |
| API URL | See "Two URL Formats" below                                                             |

Authentication is always: `Authorization: Bearer sk-YOUR-KEY`.

## Two URL Formats (Important)

ruoli supports two access protocols, and **the URL differs** — using the wrong one will fail to connect:

| Protocol                                             | API URL                                 | Typical tools                                          |
| ---------------------------------------------------- | --------------------------------------- | ------------------------------------------------------ |
| **OpenAI-compatible** (Chat Completions / Responses) | `https://ruoli.dev/v1`                  | Codex, Cherry Studio, and most OpenAI-compatible tools |
| **Native** (Anthropic / Gemini, etc.)                | `https://ruoli.dev` (**without** `/v1`) | Claude Code (Anthropic), Gemini CLI (Gemini)           |

<Warning>
  Rule of thumb: if the tool asks for an **OpenAI / OpenAI-compatible** URL, use `https://ruoli.dev/v1`; if it asks for a **native vendor** URL (Anthropic, Gemini, etc.), use `https://ruoli.dev` (without `/v1`).

  These native SDKs build the path themselves, so an extra `/v1` will break the request.
</Warning>

## Universal Configuration Example

For most OpenAI-compatible tools, these three fields are all you need:

| Field    | Value                  |
| -------- | ---------------------- |
| Provider | Custom, e.g. `ruoli`   |
| Base URL | `https://ruoli.dev/v1` |
| API Key  | `sk-YOUR-KEY`          |

Then enter the model name you want (e.g. `claude-sonnet-4-20250514`, `gpt-4o`) and you're ready to go.

## Verify the Connection

After configuring, run the command below to confirm you get a valid response (OpenAI-compatible format):

```bash theme={null}
curl https://ruoli.dev/v1/chat/completions \
  -H "Authorization: Bearer sk-YOUR-KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-20250514",
    "messages": [{"role": "user", "content": "hi"}]
  }'
```

If you get back JSON containing a `content` text field, you're connected. For common errors, see the [FAQ](/en/platform/faq).

<Info>
  **Dedicated pages for popular tools**: [Claude Code](/en/tools/claude-code) · [Codex](/en/tools/codex) · [Cherry Studio](/en/tools/cherry-studio).

  Need to manage connections for multiple CLI tools at once? Use the [CC Switch](/en/tools/cc-switch) setup helper.
</Info>