Anthropic
anthropic
API key / OAuth
Full agentic workflow — bash, files, grep, glob, agents, MCP, web search — running on Anthropic, OpenAI, Gemini, Ollama, or any of 200+ models. One CLI, zero vendor lock-in.
Providers
Configure any provider interactively with /provider — no environment variables, no config files, no restarts.
Features
No usage data, no analytics, no crash reports sent anywhere. Your prompts, code, and conversations stay entirely on your machine and go only to the provider you choose.
DuckDuckGo web search ships built-in for all providers — not just Anthropic. Optional Firecrawl integration for JS-rendered pages. No extra setup required.
Persistent memory across sessions — private and team-scoped. Automatic context pruning, token budget management, and stub eviction keep costs low as projects grow.
Bash execution, file tools, grep, glob, sub-agents, tasks, MCP servers, LSP diagnostics, Jupyter notebooks, cron scheduling, plan mode, git worktrees — all in one CLI.
Automatic cache-breakpoint management across all supported providers. Full conversation fidelity preserved — no summarization or context loss. Cache hits cut costs and latency on every turn.
Provider-aware control-center UI, launch integration, and theme support — no separate install. MCP built-in. Skills and plugins system for extensible slash commands.
Claudin originated from the Claude Code codebase and has been substantially modified to support any provider. It automatically detects your existing ~/.claude/ and migrates tokens, settings, theme, plugins, and keybindings.
Independent project. Not affiliated with, endorsed by, or sponsored by Anthropic.
Migration guide →No API key, no cost, no data sent anywhere. Pull a model with Ollama or LM Studio and start coding in under a minute.
Local setup guide →FAQ
Common questions about installing and using the Claudin coding agent CLI across different LLM providers.
Claudin AI (also called Claudin CLI) is a free, open-source terminal coding agent that brings AI-powered software development to your command line. It provides a full agentic loop — reading and writing files, running bash commands, searching the web, spawning sub-agents, and calling MCP servers — on top of any AI provider. You can power Claudin AI with Anthropic's Claude, OpenAI's GPT models, Google Gemini, DeepSeek, Mistral, or any local model via Ollama. Install with npm install -g @claudiolabs/claudin@latest.
Yes. The Claudin CLI is free to use and the source code is available on GitHub at github.com/claudio-labs/claudin. You only pay for the LLM API calls you make to whichever provider you configure — or nothing at all if you run a local model via Ollama or LM Studio. There are no subscriptions, no paywalls, and no telemetry.
Claudin is a terminal-first AI coding agent CLI that gives you a full agentic workflow — bash, files, grep, glob, sub-agents, MCP, and web search — on top of any LLM provider. Unlike Claude Code (Anthropic-only) or Cursor (IDE-only), Claudin runs in your terminal and lets you switch between Anthropic, OpenAI, Gemini, DeepSeek, Mistral, Ollama, AWS Bedrock, Google Vertex, OpenRouter, Groq and 200+ other endpoints without changing your workflow. The project originated from the Claude Code codebase and has been substantially modified to support any provider.
Claudin supports Anthropic, OpenAI, DeepSeek, Gemini, Mistral, xAI/Grok, GitHub Copilot, Codex (ChatGPT), Groq, OpenRouter, Together AI, NVIDIA NIM, MiniMax, Moonshot AI, Kimi Code, Azure OpenAI, Alibaba DashScope, Bankr, OpenCode Zen/GO, Ollama, LM Studio, Atomic Chat, AWS Bedrock, Google Vertex AI, Azure AI Foundry, and any OpenAI-compatible endpoint — over 200 providers in total. Configure any of them interactively with the /provider command — no environment variables, no config files, no restarts required.
Yes. Pull a model with Ollama or LM Studio (for example ollama pull qwen2.5-coder:7b) and launch Claudin with claudin --model qwen2.5-coder:7b. No API key, no cost, and no data leaves your machine. The full agentic toolkit — bash, files, grep, glob, web search, MCP — works the same way as with cloud providers.
Yes. On first launch, Claudin automatically detects your existing ~/.claude/ directory and migrates your tokens, settings, theme, plugins, keybindings, and memory. Just run npm install -g @claudiolabs/claudin@latest and then claudin — your existing workflow is preserved and you immediately gain the ability to switch providers.
Claudin requires Node.js 20+. Install it globally with npm install -g @claudiolabs/claudin@latest, bun install -g @claudiolabs/claudin@latest, or run it directly with npx @claudiolabs/claudin@latest. There's also a one-line installer: curl -fsSL https://claudiolabs.ai/install.sh | sh.
No. Claudin ships with zero telemetry by default. The only network calls made are to the LLM provider you explicitly configure. When you run locally with Ollama or LM Studio, nothing leaves your machine at all.