Command Code Proxy

A reverse proxy that converts Command Code API to OpenAI / Anthropic compatible endpoints. TypeScript, zero runtime dependencies.

Built by analyzing official CLI v0.32.3 network traffic to accurately replicate the Command Code API request protocol.

Features:

OpenAI Chat Completions + Anthropic Messages API
Streaming & non-streaming
Tool calling (tool_use)
Multimodal image input
Reasoning effort
Dynamic model list
Cache hit metrics
Client disconnect detection with upstream abort
Zero-output → 502 auto-retry
Consecutive timeout → 429 auto-retry
Privacy-aware logging

Quick Start

bun start        # Start (default http://0.0.0.0:3050)
bun run dev      # Watch mode (auto-reload on file changes)

API Key is passed via the Authorization request header — no need to store it in config files. Key must start with user_ (automatically matched with any prefix, e.g. Bearer token_user_xxx):

curl http://127.0.0.1:3050/v1/chat/completions \
  -H "Authorization: Bearer user_xxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek/deepseek-v4-flash","messages":[{"role":"user","content":"hi"}]}'

File Structure

commandcode/
├── config.json         # Port / log path etc.
├── LICENSE             # MIT License
├── package.json        # bun start / bun run dev
├── tsconfig.json       # TypeScript configuration
├── Dockerfile          # Container build (oven/bun:1)
├── docker-compose.yml  # Container orchestration
├── .dockerignore       # Build context exclusions
├── README.md           # This document
└── src/
    ├── server.ts       # HTTP server entry point
    ├── config.ts       # Config loader (config.json + env vars)
    ├── types.ts        # Shared TypeScript types
    ├── models.ts       # Model list (hardcoded + dynamic fetch)
    ├── logger.ts       # Logging (console + file)
    ├── helpers.ts      # Utilities (traceparent, slug, fingerprint)
    ├── http-util.ts    # Body parsing, JSON responses, auth
    ├── upstream.ts     # CC API forwarding + error mapping
    ├── runtime-state.ts # Session, version, fingerprint, timeouts
    ├── openai.ts       # OpenAI protocol conversion
    ├── anthropic.ts    # Anthropic protocol conversion
    └── routes/
        ├── chat.ts     # POST /v1/chat/completions
        ├── messages.ts # POST /v1/messages
        └── meta.ts     # GET /v1/models, GET /health

Configuration

config.json

Field	Default	Description
`port`	`3050`	Listen port
`host`	`0.0.0.0`	Listen address
`apiBase`	`https://api.commandcode.ai`	CC API base URL
`projectSlug`	`cc-proxy`	`x-project-slug` header
`logFile`	`""`	Log file path (empty = console only)
`logLevel`	`info`	Log level
`useProviderModels`	`true`	Dynamically fetch model list from Provider API
`modelRefreshIntervalMs`	`300000`	Model list cache refresh interval (5 min)

Environment Variables

Variable	Overrides
`PORT`	`port`
`HOST`	`host`
`CC_API_BASE`	`apiBase`
`PROJECT_SLUG`	`projectSlug`
`LOG_FILE`	`logFile`
`CC_USE_PROVIDER_MODELS`	`useProviderModels`

API Endpoints

`POST /v1/chat/completions`

OpenAI Chat Completions compatible. Supports streaming, non-streaming, tool calling, multimodal image input, and reasoning effort.

Request parameters:

Parameter	Required	Description
`model`	Yes	Model ID (see model list)
`messages`	Yes	Conversation messages, supports `system/user/assistant/tool` roles
`max_tokens`	No	Max tokens to generate (default 64000)
`stream`	No	SSE streaming (default false)
`temperature`	No	Sampling temperature (0-2)
`reasoning_effort`	No	Reasoning intensity: `low`/`medium`/`high`/`max`
`tools`	No	Tool definitions (OpenAI function calling format)
`tool_choice`	No	Tool selection strategy
`parallel_tool_calls`	No	Allow parallel tool calls

Simple request:

{
  "model": "deepseek/deepseek-v4-flash",
  "messages": [{ "role": "user", "content": "hello" }],
  "stream": true
}

Multimodal image input (vision model required):

{
  "model": "xiaomi/mimo-v2.5",
  "messages": [
    {
      "role": "user",
      "content": [
        { "type": "text", "text": "Describe this image" },
        { "type": "image_url", "image_url": { "url": "data:image/jpeg;base64,..." } }
      ]
    }
  ]
}

Tool calling:

{
  "model": "deepseek/deepseek-v4-flash",
  "messages": [...],
  "tools": [{
    "type": "function",
    "function": { "name": "get_weather", "description": "...", "parameters": {...} }
  }],
  "tool_choice": "auto"
}

Streaming response (SSE):

data: {"id":"chatcmpl-xxx","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"role":"assistant","reasoning_content":"thinking..."}}]}

data: {"id":"chatcmpl-xxx","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"content":"Hello"}}]}

data: {"id":"chatcmpl-xxx","object":"chat.completion.chunk","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"prompt_tokens":10,"completion_tokens":20,"total_tokens":30,"prompt_tokens_details":{"cached_tokens":8}}}

data: [DONE]

Non-streaming response (with cache hits):

{
  "id": "chatcmpl-xxx",
  "object": "chat.completion",
  "created": 1234567890,
  "model": "deepseek/deepseek-v4-flash",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello!",
        "reasoning_content": "The user said hello, I should respond."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 7558,
    "completion_tokens": 42,
    "total_tokens": 7600,
    "prompt_tokens_details": { "cached_tokens": 7552 }
  }
}

`POST /v1/messages`

Anthropic Messages API compatible endpoint. Supports streaming, non-streaming, and tool calling.

Request body:

{
  "model": "claude-sonnet-4-6",
  "max_tokens": 1000,
  "system": "You are a helpful assistant.",
  "messages": [{ "role": "user", "content": "hello" }],
  "stream": true
}

Anthropic protocol conversion (automatic):

Concept	Anthropic Format	Conversion
System prompt	Top-level `system` field	Auto-converted to OpenAI `system` message
Message content	`content` array (text/tool_use/tool_result)	Auto-mapped to corresponding roles
Tool results	`tool_result` blocks in `user` messages	Auto-converted to `role: "tool"`
Tool definitions	`input_schema`	Auto-mapped to `parameters`
`tool_choice`	`{type:"auto"/"any"/"tool"}`	`any`→`required`, `tool`→function object
Reasoning	`thinking.budget_tokens`	Auto-mapped to `reasoning_effort` (≥10000→high, ≥5000→medium, ≥2000→low)
Stop reason	`end_turn`/`max_tokens`/`tool_use`	Auto-mapped to `stop`/`length`/`tool_calls`
Token usage	`input_tokens`/`output_tokens` + cache	Passed through, cache fields mapped to Anthropic format

Streaming response (SSE, Anthropic format):

event: message_start
data: {"type":"message_start","message":{"id":"msg_xxx","type":"message","role":"assistant","content":[],"model":"...","usage":{"input_tokens":0,"output_tokens":0}}}

event: content_block_start
data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Hello"}}

event: content_block_stop
data: {"type":"content_block_stop","index":0}

event: message_delta
data: {"type":"message_delta","delta":{"stop_reason":"end_turn"},"usage":{"output_tokens":10,"cache_read_input_tokens":0,"input_tokens":100}}

event: message_stop
data: {"type":"message_stop"}

Non-streaming response:

{
  "id": "msg_xxx",
  "type": "message",
  "role": "assistant",
  "model": "deepseek/deepseek-v4-flash",
  "content": [{ "type": "text", "text": "Hello!" }],
  "stop_reason": "end_turn",
  "stop_sequence": null,
  "usage": {
    "input_tokens": 7558,
    "output_tokens": 42,
    "cache_read_input_tokens": 7552,
    "cache_creation_input_tokens": null
  }
}

`GET /v1/models`

Returns available model list. Fetched dynamically from Provider API (5 min cache), falls back to hardcoded list on failure.

`GET /health` (`/` also works)

Health check. Returns OK.

Error Codes

HTTP Status	Description
400	Invalid request format
401	API Key missing / invalid format / rejected (Key must start with `user_`)
429	Stream idle timeout (30s streaming / 90s non-streaming, SDK auto-retry, consecutive 3: reduce context hint)
502	Zero output tokens or CC upstream error
503	Service temporarily unavailable

Model List

The proxy returns a live model list via GET /v1/models. Below are common models for reference; the actual list depends on the live API response — see Command Code Pricing for plan details.

Common Models

Model ID	Name	Features
`deepseek/deepseek-v4-flash`	DeepSeek V4 Flash	Fast, general-purpose
`deepseek/deepseek-v4-pro`	DeepSeek V4 Pro	High-precision reasoning
`claude-sonnet-4-6`	Claude Sonnet 4.6	Long context
`claude-opus-4-8`	Claude Opus 4.8	Best reasoning
`claude-opus-4-7`	Claude Opus 4.7	Reasoning
`claude-haiku-4-5-20251001`	Claude Haiku 4.5	Fast, lightweight
`gpt-5.5`	GPT-5.5	OpenAI flagship
`gpt-5.4`	GPT-5.4	General-purpose
`gpt-5.4-mini`	GPT-5.4 Mini	Fast, cost-efficient
`gpt-5.3-codex`	GPT-5.3 Codex	Code generation
`moonshotai/Kimi-K2.6`	Kimi K2.6	Multimodal / frontend
`moonshotai/Kimi-K2.5`	Kimi K2.5	Multimodal / frontend
`zai-org/GLM-5.1`	GLM 5.1	Large context
`zai-org/GLM-5`	GLM 5	General-purpose
`MiniMaxAI/MiniMax-M3`	MiniMax M3	Multimodal
`MiniMaxAI/MiniMax-M2.7`	MiniMax M2.7	General-purpose
`MiniMaxAI/MiniMax-M2.5`	MiniMax M2.5	Fast, lightweight
`Qwen/Qwen3.7-Max`	Qwen 3.7 Max	Large parameters
`Qwen/Qwen3.6-Max-Preview`	Qwen 3.6 Max Preview	Preview
`Qwen/Qwen3.6-Plus`	Qwen 3.6 Plus	Balanced
`stepfun/Step-3.7-Flash`	Step 3.7 Flash	Fast reasoning
`stepfun/Step-3.5-Flash`	Step 3.5 Flash	Fast, lightweight
`xiaomi/mimo-v2.5-pro`	MiMo V2.5 Pro	Image input supported
`xiaomi/mimo-v2.5`	MiMo V2.5	Image input supported
`google/gemini-3.5-flash`	Gemini 3.5 Flash	Reasoning model
`google/gemini-3.1-flash-lite`	Gemini 3.1 Flash Lite	Fast, lightweight

⚠️ Not all models support image input. Use vision-capable models (e.g. xiaomi/mimo-v2.5, Kimi-K2.6, MiniMax-M3) for multimodal requests.

Integration Examples

Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    api_key="user_xxxxxxxxx",
    base_url="http://127.0.0.1:3050/v1",
)

response = client.chat.completions.create(
    model="deepseek/deepseek-v4-flash",
    messages=[{"role": "user", "content": "hello"}],
    stream=True,
)
for chunk in response:
    print(chunk.choices[0].delta.content or "", end="")

cURL

curl http://127.0.0.1:3050/v1/chat/completions \
  -H "Authorization: Bearer user_xxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4-flash",
    "messages": [{"role": "user", "content": "hello"}],
    "stream": true
  }'

Cursor

Add a Custom Provider in Cursor settings:

API Base URL: http://127.0.0.1:3050/v1
API Key: user_xxxxxxxxx
Model: Choose from the model list

Anthropic (Python SDK)

import anthropic

client = anthropic.Anthropic(
    api_key="user_xxxxxxxxx",
    base_url="http://127.0.0.1:3050",
)
message = client.messages.create(
    model="deepseek/deepseek-v4-flash",
    max_tokens=1000,
    system="You are helpful.",
    messages=[{"role": "user", "content": "hello"}],
)
print(message.content[0].text)

OpenCode

{
  "provider": "openai-compatible",
  "baseUrl": "http://127.0.0.1:3050/v1",
  "apiKey": "user_xxxxxxxxx"
}

Anti-Detection

Based on analysis of official CLI v0.32.3 traffic:

Mechanism	Implementation
Per-Key Session	One session per API key, 12h expiry + 1h random jitter
Session ID	`x-session-id` header, per-key with automatic rotation
Version	`x-command-code-version` auto-fetched from npm registry (24h refresh)
CLI Envelope	config/memory/taste/permissionMode/params/threadId
OpenTelemetry	`traceparent` (W3C Trace Context)
Environment	`x-cli-environment: production`
Project Slug	Fake `x-project-slug` derived from session ID (matches real CLI slug format)
Device Fingerprint	Generated CPU/memory/timezone/MAC fingerprint per key (mimics real CLI)
Reasoning Effort	`reasoning_effort` pass-through (low/medium/high/max)
Key Validation	Regex `user_[a-zA-Z0-9_-]+`, auto-cleans extra paths/prefixes, rejects `sk-xxx` format
Stream Timeout	30s streaming / 90s non-streaming → 429 with SDK auto-retry
Consecutive Timeout	3 consecutive timeouts before "reduce context" hint
Zero-Output Guard	outputTokens=0 → 502 error (SDK auto-retry, anti false billing)
Upstream Abort	`AbortController` on client disconnect + all error paths
Privacy Logging	No API key fragments, no error bodies, no stack traces in logs

Protocol Details

CC API Request Structure

{
  "config": {
    "workingDir": "C:\\project",
    "date": "2026-06-07",
    "environment": "win32-x64, Node.js v24.16.0",
    "structure": [],
    "isGitRepo": false,
    "currentBranch": "",
    "mainBranch": "",
    "gitStatus": "",
    "recentCommits": []
  },
  "memory": "",
  "taste": "",
  "skills": "",
  "permissionMode": "standard",
  "params": {
    "model": "deepseek/deepseek-v4-flash",
    "messages": [...],
    "max_tokens": 64000,
    "stream": true,
    "reasoning_effort": "max"
  },
  "threadId": "<uuid>"
}

CC API Image Message Format

The CLI sends images in this format:

{
  "role": "user",
  "content": [
    { "type": "image", "image": "data:image/jpeg;base64,..." },
    { "type": "text", "text": "What does this image say?" }
  ]
}

The proxy receives OpenAI image_url format and converts it to the above CC format transparently.

Docker Deployment

Quick Start (docker compose)

docker compose up -d

The proxy will listen on http://0.0.0.0:3050. Set PROXY_PORT to customize the host port:

PROXY_PORT=13050 docker compose up -d

Build from Source

docker build -t commandcode-proxy:latest .
docker run -d -p 3050:3050 -e PORT=3050 commandcode-proxy:latest

Multi-Architecture Build

bun run docker:build:multi

Environment Variables

Variable	Default	Description
`PORT`	`3050`	Container listen port
`PROXY_PORT`	`3050`	Host port (compose only)

Disclaimer

This project is for educational and research purposes only.

Unofficial: This project is not affiliated with Command Code in any way.
Personal Use: Users assume all responsibility. Please comply with the Command Code Terms of Service.
API Key: This project does not collect, upload, or leak your API Key. The key must be sent in every request via the Authorization: Bearer <key> header and is never stored in configuration.
Compliance: The protocol is based on passive observation of local CLI network traffic. No unauthorized access, cracking, or tampering of the server has been performed.
Account Risk: Keep usage frequency consistent with normal CLI usage. Extremely high concurrent calls may trigger risk controls.

Development

bun run dev          # Watch mode (auto-reload on file changes)
bun run typecheck    # Type-check without emitting
bun run lint         # ESLint
bun run format       # Prettier

Name		Name	Last commit message	Last commit date
examples		examples
src		src
.dockerignore		.dockerignore
.gitignore		.gitignore
.prettierrc		.prettierrc
AGENTS.md		AGENTS.md
CLAUDE.md		CLAUDE.md
Dockerfile		Dockerfile
LICENSE		LICENSE
README.md		README.md
bun.lock		bun.lock
eslint.config.mjs		eslint.config.mjs
package.json		package.json
skills-lock.json		skills-lock.json
tsconfig.json		tsconfig.json

Navigation Menu

Repository files navigation

Command Code Proxy

Quick Start

File Structure

Configuration

config.json

Environment Variables

API Endpoints

`POST /v1/chat/completions`

`POST /v1/messages`

`GET /v1/models`

`GET /health` (`/` also works)

Error Codes

Model List

Common Models

Integration Examples

Python (OpenAI SDK)

cURL

Cursor

Anthropic (Python SDK)

OpenCode

Anti-Detection

Protocol Details

CC API Request Structure

CC API Image Message Format

Docker Deployment

Quick Start (docker compose)

Build from Source

Multi-Architecture Build

Environment Variables

Disclaimer

Development

About

Releases

Packages

Contributors

Languages

Latest commit

History

Folders and files

Repository files navigation

Command Code Proxy

Quick Start

File Structure

Configuration

config.json

Environment Variables

API Endpoints

POST /v1/chat/completions

POST /v1/messages

GET /v1/models

GET /health (/ also works)

Error Codes

Model List

Common Models

Integration Examples

Python (OpenAI SDK)

cURL

Cursor

Anthropic (Python SDK)

OpenCode

Anti-Detection

Protocol Details

CC API Request Structure

CC API Image Message Format

Docker Deployment

Quick Start (docker compose)

Build from Source

Multi-Architecture Build

Environment Variables

Disclaimer

Development

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages

`POST /v1/chat/completions`

`POST /v1/messages`

`GET /v1/models`

`GET /health` (`/` also works)