Client Setup Overview

ClawRouter works with any AI client that supports a custom base URL. This page explains the two setup methods and what all clients have in common -- then points you to the dedicated guide for your client.

Version 1.0.18

Dedicated Client Guides

Step-by-step instructions -- exact config files, copy-paste snippets, and troubleshooting -- for each client:

ClientGuide
OpenClawClient Setup > OpenClaw
OpenCodeClient Setup > OpenCode
Claude CodeClient Setup > Claude Code
Codex CLIClient Setup > Codex CLI
Qwen CodeClient Setup > Qwen Code
DeepSeek HarnessClient Setup > DeepSeek Harness
Any other client (Cline, Aider, Kilo CLI, ...)Client Setup > Other / Custom Clients

What Every Client Needs

Regardless of the client, three things are always required:

  1. A running ClawRouter -- clawrouter status should show the service up (default port 3030).
  2. A configured provider -- created in the dashboard, with API keys added (unless it is a keyless preset).
  3. The proxy API key -- shown in Settings > Proxy API Key. Clients authenticate to ClawRouter with this key; ClawRouter then injects the real upstream key itself.

Requests without a valid proxy API key get HTTP 401. The key can be sent as Authorization: Bearer <key> (OpenAI style) or x-api-key: <key> (Anthropic style).

Any client, any provider: ClawRouter detects the client's API format from the request path and translates to the provider's format automatically -- every client below works with every provider, regardless of format.

Every provider page has a "Prompt for AI" button (on the Base URL banner) that opens a tabbed dialog with ready-to-paste setup instructions for 7 AI clients:

TabTarget Client
OpenClaw (default)openclaw.json via config.patch, using native provider IDs where available
OpenCodeopencode.json custom provider (@ai-sdk/openai-compatible, @ai-sdk/openai, or @ai-sdk/anthropic)
Claude Code~/.claude/settings.json env block
Codex CLI~/.codex/config.toml model provider (Responses API)
Qwen CodeEnv vars or ~/.qwen/settings.json modelProviders
DeepSeek Harness~/.dsh/settings.yaml llm-pi-ai provider
Custom / OtherGeneric endpoint reference, env export blocks, and a curl example
  1. Open the provider's detail page.
  2. Click "Prompt for AI" on the Base URL banner at the top.
  3. Select your client's tab, click Copy, and paste the prompt to your AI agent (or follow the instructions yourself).

How templates are generated:

  • Model lists come from the provider's actual saved models (Models tab). If none are saved, the preset's recommended models are used.
  • The apiKey is your real proxy API key (from Settings > Proxy API Key), embedded automatically -- no manual copying needed.
  • Every prompt includes a Model Parameters section: the client's exact config keys for context-window / max-output / reasoning (verified against current client docs), known reasoning/thinking facts for the provider (e.g. Qwen's enable_thinking + thinking_budget, or reasoning_effort low/medium/xhigh on qwen3.8-max), and a link to the provider's official model docs so the agent fills in exact per-model values instead of guessing.
  • All tabs work with every provider. When the client's native API format differs from the provider's, the tab shows an amber note that ClawRouter translates automatically. Reasoning/thinking parameters are translated into the provider's native dialect too.

Method 2: Manual Configuration

Prefer editing config files yourself? Each dedicated client guide has the exact file path and a copy-paste config block. The only client-specific differences are the config file location and the base URL suffix:

ClientBase URL Suffix
OpenClaw/v1 (or /v1beta for Google) -- as shown on the provider's banner
OpenCode/v1 required
Codex CLI/v1 required
Qwen Code/v1 required
DeepSeek Harness/v1 required
Claude CodeNo /v1 -- Claude Code appends /v1/messages itself

The base URL always starts with http://localhost:3030/proxy/{provider-id} (adjust host/port if you changed them).