Open Source ClaudeCode Leaked. Same functionalities as Claude Code by Anthropic. Written in Rust, Ready to use immediately for free.
See the codeA high-performance, open-source AI coding agent written in Rust.
ClaudeCode is a terminal-native CLI agent designed to bring advanced LLM capabilities directly into your development workflow. Built for speed, safety, and efficiency, it provides an interactive agent shell, workspace-aware tools, and persistent session management. It is an independent open-source implementation inspired by Claude Code, not the official Anthropic product.
ratatui interface showing the agent's pipeline, a workspace tree that recolors as files are scanned/read/modified, and a syntax-highlighted diff that streams in character-by-character.From the repository root:
cd rust
cargo build --release -p claw-cli
# Install locally to your PATH for global access
cargo install --path crates/claw-cli --locked
Start the interactive shell:
claw
Run a single prompt:
claw prompt "summarize this workspace"
Resume a previous session:
claw --resume session.json /status
Launch the live split-pane dashboard (currently runs a scripted demo session):
claw dashboard # alias: claw tui
The dashboard renders three panels — Agent Pipeline (left, a live step log), Workspace + System State (right, a file tree that recolors as files are scanned/read/modified, plus a progress gauge), and Live Diff (bottom, a syntax-highlighted diff that types in line-by-line). Press q, Esc, or Ctrl-C to exit. The UI is driven entirely by an AgentEvent stream, so it can be wired to the real agent loop without touching the rendering code.
Run claw --help for the full command list, including agents, skills, and system-prompt flows.
Configure your environment variables based on your preferred provider:
export ANTHROPIC_API_KEY="..."
export ANTHROPIC_BASE_URL="https://api.anthropic.com"
export OPENAI_API_KEY="..."
export OPENAI_BASE_URL="https://api.openai.com/v1"
export XAI_API_KEY="..."
export XAI_BASE_URL="https://api.x.ai"
You can also authenticate via the CLI:
claw login
What is this project? This is an independent, open-source implementation of a terminal-based coding agent, architecturally inspired by Claude Code.
Why Rust? Rust provides the performance, concurrency, and memory safety required for a tool that interacts deeply with local file systems and high-latency LLM APIs.
Can I use local models? Yes, if your local inference server exposes an OpenAI-compatible API and you point the relevant base URL and API key at it.
Is this the official Anthropic Claude Code? No, this is a community-driven open-source project.
We welcome contributions. Please refer to CLAW.md for workspace-specific workflow guidance.
cd rust
cargo fmt
cargo clippy --workspace --all-targets -- -D warnings
cargo test --workspace
rust/: Core CLI and runtime implementation.src/: Python support code and utilities.tests/: Verification suites for agentic behaviors.CLAW.md: Internal workflow documentation.1,003 followers · starred Apr 2026
61 followers · starred Apr 2026
69 followers · starred Jun 2026
15 followers · starred Oct 2026
Open Source ClaudeCode Leaked. Same functionalities as Claude Code by Anthropic. Written in Rust, Ready to use immediately for free.
See the codeA high-performance, open-source AI coding agent written in Rust.
ClaudeCode is a terminal-native CLI agent designed to bring advanced LLM capabilities directly into your development workflow. Built for speed, safety, and efficiency, it provides an interactive agent shell, workspace-aware tools, and persistent session management. It is an independent open-source implementation inspired by Claude Code, not the official Anthropic product.
ratatui interface showing the agent's pipeline, a workspace tree that recolors as files are scanned/read/modified, and a syntax-highlighted diff that streams in character-by-character.From the repository root:
cd rust
cargo build --release -p claw-cli
# Install locally to your PATH for global access
cargo install --path crates/claw-cli --locked
Start the interactive shell:
claw
Run a single prompt:
claw prompt "summarize this workspace"
Resume a previous session:
claw --resume session.json /status
Launch the live split-pane dashboard (currently runs a scripted demo session):
claw dashboard # alias: claw tui
The dashboard renders three panels — Agent Pipeline (left, a live step log), Workspace + System State (right, a file tree that recolors as files are scanned/read/modified, plus a progress gauge), and Live Diff (bottom, a syntax-highlighted diff that types in line-by-line). Press q, Esc, or Ctrl-C to exit. The UI is driven entirely by an AgentEvent stream, so it can be wired to the real agent loop without touching the rendering code.
Run claw --help for the full command list, including agents, skills, and system-prompt flows.
Configure your environment variables based on your preferred provider:
export ANTHROPIC_API_KEY="..."
export ANTHROPIC_BASE_URL="https://api.anthropic.com"
export OPENAI_API_KEY="..."
export OPENAI_BASE_URL="https://api.openai.com/v1"
export XAI_API_KEY="..."
export XAI_BASE_URL="https://api.x.ai"
You can also authenticate via the CLI:
claw login
What is this project? This is an independent, open-source implementation of a terminal-based coding agent, architecturally inspired by Claude Code.
Why Rust? Rust provides the performance, concurrency, and memory safety required for a tool that interacts deeply with local file systems and high-latency LLM APIs.
Can I use local models? Yes, if your local inference server exposes an OpenAI-compatible API and you point the relevant base URL and API key at it.
Is this the official Anthropic Claude Code? No, this is a community-driven open-source project.
We welcome contributions. Please refer to CLAW.md for workspace-specific workflow guidance.
cd rust
cargo fmt
cargo clippy --workspace --all-targets -- -D warnings
cargo test --workspace
rust/: Core CLI and runtime implementation.src/: Python support code and utilities.tests/: Verification suites for agentic behaviors.CLAW.md: Internal workflow documentation.1,003 followers · starred Apr 2026
61 followers · starred Apr 2026
69 followers · starred Jun 2026
15 followers · starred Oct 2026