OpenCLI

A coding agent that runs locally in your terminal and works with any OpenAI-compatible API - a hosted provider, a self-hosted gateway, or a model running on your own machine.

This build ships no models and no API keys. You declare the providers and models you want in ~/.opencli/config.toml, so the binary carries no opinion about where your inference comes from and no secrets.

---

Install

Desktop app - opencli.ai/download, or straight from the latest release:

Platform	File
macOS, Apple Silicon	OpenCLI-macos-aarch64.dmg
macOS, Intel	OpenCLI-macos-x86_64.dmg
Windows	OpenCLI-windows-x86_64-setup.exe
Linux	OpenCLI-linux-x86_64.AppImage

The app updates itself. The first launch will not go quietly, though: these builds are not signed with an Apple or Microsoft certificate, so the operating system cannot tell who made them. On macOS, run xattr -dr com.apple.quarantine /Applications/OpenCLI.app once, or right-click the app and choose Open. On Windows, SmartScreen wants More info, then Run anyway.

Command line - one binary, macOS and Linux:

curl -fsSL https://opencli.ai/install.sh | sh

or with npm, anywhere Node runs:

npm install -g @ai-dashboad/opencli

---

Quickstart

Build from source

Requires a recent Rust toolchain.

git clone https://github.com/ai-dashboad/opencli.git
cd opencli/opencli-rs
cargo build --release -p opencli-cli --bin opencli

The binary lands at opencli-rs/target/release/opencli; put it somewhere on your PATH.

Configure a model

Create ~/.opencli/config.toml pointing at any OpenAI-compatible endpoint:

model = "my-model"

[model_providers.my-gateway]
name = "My Gateway"
base_url = "https://gateway.example.com/v1"
env_key = "MY_GATEWAY_API_KEY"
wire_api = "chat"

[[models]]
model = "my-model"
provider = "my-gateway"
display_name = "My Model"
context_window = 128000
reasoning_efforts = ["low", "medium", "high"]

Then:

export MY_GATEWAY_API_KEY=...
opencli

Models declared this way appear in the /model picker. Local runtimes work the same way - built-in ollama and lmstudio providers point at localhost.

For OpenAI itself, set OPENAI_API_KEY and use the built-in openai provider.

See docs/config.md for the full reference.

Docs

• Configuration - providers, models, routing, sandboxing
• Contributing
• Installing & building

License and attribution

Licensed under the Apache-2.0 License.

This project is derived from OpenAI OpenCLI, Copyright 2025 OpenAI, also licensed under Apache-2.0. See NOTICE for full attribution.