Package Exports
- api-switch-cc
This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (api-switch-cc) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.
Readme
api-switch-cc
Claude Code 多模型代理 — 一键切换 DeepSeek、Qwen、GLM、SenseNova、NVIDIA 等大模型。协议自动转换,零配置即用。 LLM API proxy for Claude Code — route to DeepSeek, Qwen, GLM, SenseNova, NVIDIA and more with automatic protocol conversion.
npm install -g api-switch-cc
api-switch setup deepseek --key sk-xxx
api-switch provider known # 查看 13 个内置厂商
api-switch use deepseek-chat
api-switch serveUsage
# Install globally
npm install -g api-switch-cc
# Or run directly
npx api-switch-cc setup agnes --key sk-xxx
npx api-switch-cc serveFeatures
- 13 built-in provider presets — DeepSeek, Qwen, GLM, SenseNova (商汤), NVIDIA (英伟达), Moonshot, Agnes AI (free), Kimi, Yi, StepFun, ERNIE, Hunyuan, APIFree
- Security hardening — Encrypted API key storage (AES-256-GCM), Bearer token auth, rate limiting, TLS support
- Auto sync settings.json — serve startup automatically syncs Claude Code base URL
- Port persistence —
serve -p <port>saves to config;--no-save-portto skip - Auto-update with verification — checks for new versions, SHA256 checksum validation
- Graceful shutdown — SIGINT/SIGTERM signal handling
- Protocol conversion — Anthropic ↔ OpenAI bidirectional (SSE streaming, tool calls, multimodal images)
- Multimodal support — Full image conversion (base64 and URL) for Vision models
- DeepSeek compatible — auto-strips orphaned tool_calls for strict providers
- Hot-reload config — changes take effect without restart
- Web dashboard — real-time request monitoring at
/admin/(localhost only) - Usage tracking — daily token statistics with cache hit rates
- Context retention — tested across all configured models
Quick Start
# 1. Install
npm install -g api-switch-cc
# 2. See available providers
api-switch provider known
# 3. Add a provider
api-switch provider add deepseek --key sk-xxx
# 4. Import models
api-switch model import deepseek
# 5. Switch model & start
api-switch use deepseek-chat
api-switch serve # default :8080, auto-syncs settings.jsonCommands
| Command | Description |
|---|---|
api-switch serve |
Start proxy (default :8080) |
api-switch serve -p 9000 |
Start on custom port, persist to config |
api-switch use <model> |
Switch active model in Claude Code |
api-switch provider known |
List 13 built-in providers |
api-switch provider list |
List configured providers |
api-switch provider add |
Add provider (supports presets) |
api-switch doctor |
Run diagnostics |
api-switch monitor |
View real-time traffic |
api-switch usage |
Token usage statistics |
api-switch update |
Update to latest version |
Security Configuration
# ~/.api-switch.yaml
server:
port: 8080
auth_token: "" # Set to require Bearer authentication
rate_limit: 100 # Requests per minute per IP (0 = disabled)
tls_cert: "" # TLS certificate path (optional)
tls_key: "" # TLS key path (optional)Security features:
- Encrypted API keys — Stored with AES-256-GCM, auto-migrates from plaintext
- Bearer authentication — Set
auth_tokento requireAuthorization: Bearer <token> - Rate limiting — Protect against DoS abuse
- TLS/HTTPS — Enable encrypted communication
- Update verification — SHA256 checksum validation for auto-updates
See the full documentation on GitHub.
Supported platforms
- Linux (amd64, arm64)
- macOS (amd64, arm64)
- Windows (amd64)