Use Claude Code with any LLM — not just Claude.
OpenClaude is a fork of the Claude Code source leak (exposed via npm source maps on March 31, 2026). We added an OpenAI-compatible provider shim so you can plug in GPT-4o, DeepSeek, Gemini, Llama, Mistral, or any model that speaks the OpenAI chat completions API. It now also supports the ChatGPT Codex backend for codexplan and codexspark.
All of Claude Code's tools work — bash, file read/write/edit, grep, glob, agents, tasks, MCP — just powered by whatever model you choose.
npm install -g @gitlawb/openclaude# Clone from gitlawb
git clone https://node.gitlawb.com/z6MkqDnb7Siv3Cwj7pGJq4T5EsUisECqR8KpnDLwcaZq5TPr/openclaude.git
cd openclaude
# Install dependencies
bun install
# Build
bun run build
# Link globally (optional)
npm linkgit clone https://node.gitlawb.com/z6MkqDnb7Siv3Cwj7pGJq4T5EsUisECqR8KpnDLwcaZq5TPr/openclaude.git
cd openclaude
bun install
bun run devexport CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=sk-your-key-here
export OPENAI_MODEL=gpt-4o# If installed via npm
openclaude
# If built from source
bun run dev
# or after build:
node dist/cli.mjsThat's it. The tool system, streaming, file editing, multi-step reasoning — everything works through the model you picked.
The npm package name is @gitlawb/openclaude, but the installed CLI command is still openclaude.
export CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=sk-...
export OPENAI_MODEL=gpt-4ocodexplan maps to GPT-5.4 on the Codex backend with high reasoning.
codexspark maps to GPT-5.3 Codex Spark for faster loops.
If you already use the Codex CLI, OpenClaude will read ~/.codex/auth.json
automatically. You can also point it elsewhere with CODEX_AUTH_JSON_PATH or
override the token directly with CODEX_API_KEY.
export CLAUDE_CODE_USE_OPENAI=1
export OPENAI_MODEL=codexplan
# optional if you do not already have ~/.codex/auth.jsonexport CODEX_API_KEY=...
openclaudeexport CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=sk-...
export OPENAI_BASE_URL=https://api.deepseek.com/v1
export OPENAI_MODEL=deepseek-chatexport CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=sk-or-...
export OPENAI_BASE_URL=https://openrouter.ai/api/v1
export OPENAI_MODEL=google/gemini-2.0-flashollama pull llama3.3:70b
export CLAUDE_CODE_USE_OPENAI=1
export OPENAI_BASE_URL=http://localhost:11434/v1
export OPENAI_MODEL=llama3.3:70b
# no API key needed for local modelsexport CLAUDE_CODE_USE_OPENAI=1
export OPENAI_BASE_URL=http://localhost:1234/v1
export OPENAI_MODEL=your-model-nameexport CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=...
export OPENAI_BASE_URL=https://api.together.xyz/v1
export OPENAI_MODEL=meta-llama/Llama-3.3-70B-Instruct-Turboexport CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=gsk_...
export OPENAI_BASE_URL=https://api.groq.com/openai/v1
export OPENAI_MODEL=llama-3.3-70b-versatileexport CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=...
export OPENAI_BASE_URL=https://api.mistral.ai/v1
export OPENAI_MODEL=mistral-large-latestexport CLAUDE_CODE_USE_OPENAI=1
export OPENAI_API_KEY=your-azure-key
export OPENAI_BASE_URL=https://your-resource.openai.azure.com/openai/deployments/your-deployment/v1
export OPENAI_MODEL=gpt-4o| Variable | Required | Description |
|---|---|---|
CLAUDE_CODE_USE_OPENAI | Yes | Set to 1 to enable the OpenAI provider |
OPENAI_API_KEY | Yes* | Your API key (*not needed for local models like Ollama) |
OPENAI_MODEL | Yes | Model name (e.g. gpt-4o, deepseek-chat, llama3.3:70b) |
OPENAI_BASE_URL | No | API endpoint (defaults to https://api.openai.com/v1) |
CODEX_API_KEY | Codex only | Codex/ChatGPT access token override |
CODEX_AUTH_JSON_PATH | Codex only | Path to a Codex CLI auth.json file |
CODEX_HOME | Codex only | Alternative Codex home directory (auth.json will be read from here) |
OPENCLAUDE_DISABLE_CO_AUTHORED_BY | No | Set to 1 to suppress the default Co-Authored-By trailer in generated git commit messages |
You can also use ANTHROPIC_MODEL to override the model name. OPENAI_MODEL takes priority.
OpenClaude PR bodies use OpenClaude branding by default. OPENCLAUDE_DISABLE_CO_AUTHORED_BY only affects the commit trailer, not PR attribution text.
Use these commands to keep the CLI stable and catch environment mistakes early:
# quick startup sanity check
bun run smoke
# validate provider env + reachability
bun run doctor:runtime
# print machine-readable runtime diagnostics
bun run doctor:runtime:json
# persist a diagnostics report to reports/doctor-runtime.json
bun run doctor:report
# full local hardening check (smoke + runtime doctor)
bun run hardening:check
# strict hardening (includes project-wide typecheck)
bun run hardening:strictNotes:
doctor:runtimefails fast ifCLAUDE_CODE_USE_OPENAI=1with a placeholder key (SUA_CHAVE) or a missing key for non-local providers.- Local providers (for example
http://localhost:11434/v1) can run withoutOPENAI_API_KEY. - Codex profiles validate
CODEX_API_KEYor the Codex CLI auth file and probePOST /responsesinstead ofGET /models.
Use profile launchers to avoid repeated environment setup:
# one-time profile bootstrap (prefer viable local Ollama, otherwise OpenAI)
bun run profile:init
# preview the best provider/model for your goal
bun run profile:recommend -- --goal coding --benchmark
# auto-apply the best available local/openai provider/model for your goal
bun run profile:auto -- --goal latency
# codex bootstrap (defaults to codexplan and ~/.codex/auth.json)
bun run profile:codex
# openai bootstrap with explicit key
bun run profile:init -- --provider openai --api-key sk-...
# ollama bootstrap with custom model
bun run profile:init -- --provider ollama --model llama3.1:8b
# ollama bootstrap with intelligent model auto-selection
bun run profile:init -- --provider ollama --goal coding
# codex bootstrap with a fast model alias
bun run profile:init -- --provider codex --model codexspark
# launch using persisted profile (.openclaude-profile.json)
bun run dev:profile
# codex profile (uses CODEX_API_KEY or ~/.codex/auth.json)
bun run dev:codex
# OpenAI profile (requires OPENAI_API_KEY in your shell)
bun run dev:openai
# Ollama profile (defaults: localhost:11434, llama3.1:8b)
bun run dev:ollamaprofile:recommend ranks installed Ollama models for latency, balanced, or coding, and profile:auto can persist the recommendation directly.
If no profile exists yet, dev:profile now uses the same goal-aware defaults when picking the initial model.
Use --provider ollama when you want a local-only path. Auto mode falls back to OpenAI when no viable local chat model is installed.
Goal-based Ollama selection only recommends among models that are already installed and reachable from Ollama.
Use profile:codex or --provider codex when you want the ChatGPT Codex backend.
dev:openai, dev:ollama, and dev:codex run doctor:runtime first and only launch the app if checks pass.
For dev:ollama, make sure Ollama is running locally before launch.
- All tools: Bash, FileRead, FileWrite, FileEdit, Glob, Grep, WebFetch, WebSearch, Agent, MCP, LSP, NotebookEdit, Tasks
- Streaming: Real-time token streaming
- Tool calling: Multi-step tool chains (the model calls tools, gets results, continues)
- Images: Base64 and URL images passed to vision models
- Slash commands: /commit, /review, /compact, /diff, /doctor, etc.
- Sub-agents: AgentTool spawns sub-agents using the same provider
- Memory: Persistent memory system
- No thinking mode: Anthropic's extended thinking is disabled (OpenAI models use different reasoning)
- No prompt caching: Anthropic-specific cache headers are skipped
- No beta features: Anthropic-specific beta headers are ignored
- Token limits: Defaults to 32K max output — some models may cap lower, which is handled gracefully
The shim (src/services/api/openaiShim.ts) sits between Claude Code and the LLM API:
Claude Code Tool System
|
v
Anthropic SDK interface (duck-typed)
|
v
openaiShim.ts <-- translates formats
|
v
OpenAI Chat Completions API
|
v
Any compatible model
It translates:
- Anthropic message blocks → OpenAI messages
- Anthropic tool_use/tool_result → OpenAI function calls
- OpenAI SSE streaming → Anthropic stream events
- Anthropic system prompt arrays → OpenAI system messages
The rest of Claude Code doesn't know it's talking to a different model.
Not all models are equal at agentic tool use. Here's a rough guide:
| Model | Tool Calling | Code Quality | Speed |
|---|---|---|---|
| GPT-4o | Excellent | Excellent | Fast |
| DeepSeek-V3 | Great | Great | Fast |
| Gemini 2.0 Flash | Great | Good | Very Fast |
| Llama 3.3 70B | Good | Good | Medium |
| Mistral Large | Good | Good | Fast |
| GPT-4o-mini | Good | Good | Very Fast |
| Qwen 2.5 72B | Good | Good | Medium |
| Smaller models (<7B) | Limited | Limited | Very Fast |
For best results, use models with strong function/tool calling support.
src/services/api/openaiShim.ts — NEW: OpenAI-compatible API shim (724 lines)
src/services/api/client.ts — Routes to shim when CLAUDE_CODE_USE_OPENAI=1
src/utils/model/providers.ts — Added 'openai' provider type
src/utils/model/configs.ts — Added openai model mappings
src/utils/model/model.ts — Respects OPENAI_MODEL for defaults
src/utils/auth.ts — Recognizes OpenAI as valid 3P provider
6 files changed. 786 lines added. Zero dependencies added.
This is a fork of instructkr/claude-code, which mirrored the Claude Code source snapshot that became publicly accessible through an npm source map exposure on March 31, 2026.
The original Claude Code source is the property of Anthropic. This repository is not affiliated with or endorsed by Anthropic.
This repository is provided for educational and research purposes. The original source code is subject to Anthropic's terms. The OpenAI shim additions are public domain.