Documentation Index
Fetch the complete documentation index at: https://docs.ccs.kaitran.ca/llms.txt
Use this file to discover all available pages before exploring further.
Qwen Provider
Access Qwen Code models (Qwen3 Coder) through OAuth authentication with Device Code flow. No API keys required - authenticate via browser and use instantly.Overview
The Qwen provider enables Claude Code to use Qwen’s coding-specialized models through OAuth authentication. Uses Device Code flow for enhanced security. Provider Identifier:qwen
Command: ccs qwen
Supported Models (7 total):
- Qwen3 Max (flagship, 256K context) - Claude Opus equivalent
- Qwen3 Coder Plus (balanced, 1M context) - Claude Sonnet equivalent
- Qwen3 Coder Flash (fast, optimized for speed) - Claude Haiku equivalent
- Qwen3 Max Preview (experimental variant)
- Qwen3 235B (specialized variant)
- Qwen3 VL Plus (vision-language model)
- Qwen3 32B (lightweight variant)
Model Specifications
Tier Equivalence
Qwen models map to Claude tier equivalence for preset configuration:| Qwen Model | Claude Tier | Context | Use Case |
|---|---|---|---|
| Qwen3 Max | Opus | 256K | Complex reasoning, long documents |
| Qwen3 Coder Plus | Sonnet | 1M | Balanced code & analysis, extended context |
| Qwen3 Coder Flash | Haiku | Fast | Quick code generation, API responses |
Extended Context Support
- Qwen3 Coder Plus: Supports 1M token context window
- Qwen3 Max: Supports 256K context (substantial for most tasks)
- Use
--1mflag to enable extended context when available
Model Selection
Qwen provider does not support interactive model configuration via
--config. Use environment variable override or manual settings file editing.Authentication
First-Time Setup
- CCS displays device code and verification URL
- User opens URL in any browser
- Enter device code to authorize
- Token automatically cached when complete
- No callback server required (more secure)
Manual Authentication
Logout
Multi-Account Support
Manage multiple Qwen accounts for different projects:Add Account
List Accounts
Switch Account
Rename Account
~/.ccs/cliproxy/accounts.json
Auto-Nicknames: Generated from email prefix (e.g., john@example.com → john)
Configuration Options
Model Settings
Qwen provider settings stored in: Path:~/.ccs/qwen.settings.json
Configuration via Dashboard:
CLIProxy Integration
Qwen integrates with unified CLIProxy config: Path:~/.ccs/cliproxy/config.yaml
Provider Routing: Requests routed to /api/provider/qwen endpoint
Concurrent Usage: Run Qwen + other providers simultaneously without conflicts
Commands
Basic Usage
Authentication Flags
| Flag | Description |
|---|---|
--auth | Trigger OAuth flow (no session) |
--logout | Clear all tokens |
--auth --add | Add new account (keep existing) |
Account Management Flags
| Flag | Description |
|---|---|
--accounts | List all accounts |
--use <nickname> | Switch default account |
--nickname <name> | Rename default account |
Storage Locations
| Path | Purpose |
|---|---|
~/.ccs/cliproxy/auth/qwen-*.json | OAuth tokens (one per account) |
~/.ccs/cliproxy/accounts.json | Account registry with nicknames |
~/.ccs/qwen.settings.json | Provider-specific settings |
~/.ccs/cliproxy/config.yaml | CLIProxy unified config |
Token Management
Storage Format:Device Code Flow Details
Why Device Code?- More secure (no callback server listening on local port)
- Works seamlessly in headless environments
- No port conflicts or firewall issues
- User completes auth on trusted device
- CCS requests device code from Qwen API
- Displays code + URL to user
- User authorizes in browser (any device)
- CCS polls Qwen API for token availability
- Token saved when authorization complete
null (no server required)
Requirements
- Qwen subscription
- Browser for OAuth (on any device - not necessarily same machine)
- Network access to Qwen authentication servers
Troubleshooting
Device Code Expired
Error: “Device code expired before authorization” Solution: Device codes expire after 10-15 minutes. Re-runccs qwen --auth to get new code.
Authorization Not Detected
Multi-Account Issues
Token Expiry
Tokens refresh automatically via CLIProxyAPI. If refresh fails:Model Alias Normalization
Available since v7.49.0
qwen3-max, Qwen3-Max, and qwen3_max all resolve to the same model. This prevents 406 errors from case or format mismatches in config files.
Cost Information
Qwen Subscription Required: Check qwen.ai for current pricing Usage Limits: Subject to Qwen’s rate limits and subscription tierNext Steps
Multi-Account Setup
Manage multiple Qwen accounts
Dashboard Config
Configure via web interface
