OpenJarvis/docs/getting-started/install.md

Installation

OpenJarvis ships a one-line installer for macOS, Linux, and WSL2.

curl -fsSL https://openjarvis.ai/install.sh | bash

!!! warning "openjarvis.ai SSL fallback (issue #337)" If curl fails on openjarvis.ai with sslv3 alert handshake failure, fetch the same script from the GitHub mirror until the domain is restored:

```bash
curl -fsSL https://raw.githubusercontent.com/open-jarvis/OpenJarvis/main/scripts/install/install.sh | bash
```

The installer itself pulls everything else (uv, the project source,
Ollama) from independent CDNs (`astral.sh`, `github.com`,
`ollama.com`), so the rest of install proceeds normally.

About 3 minutes on a typical broadband connection. Type jarvis to start chatting.

What the installer does

Phase	Step	Where
Foreground	Install uv (Python package manager)	~/.cargo/bin/ or ~/.local/bin/
Foreground	Clone OpenJarvis repo	~/.openjarvis/src/
Foreground	Create Python 3.11 venv	~/.openjarvis/.venv/
Foreground	uv pip install -e . (editable install)	venv
Foreground	Install Ollama	system default
Foreground	Start ollama serve	systemd-user / launchd / nohup
Foreground	Pull qwen3.5:2b (~1.5 GB)	Ollama's model store
Foreground	Write config.toml (auto-detected hardware + engine + model)	~/.openjarvis/config.toml
Foreground	Symlink jarvis and jarvis-uninstall	~/.local/bin/
Foreground	Add ~/.local/bin to PATH if missing (with on-screen notice)	~/.bashrc or ~/.zshrc
Background	Install Rust toolchain via rustup	~/.cargo/
Background	Build the maturin extension (memory + security features)	venv
Background	Pull hardware-tier and tier+1 models	Ollama's model store

What the installer does NOT touch

• Your existing Python installations
• Your ~/.bashrc / ~/.zshrc other than appending one PATH line (with on-screen notice)
• Your existing Ollama models
• Any other tool or dotfile

Idempotent re-runs

Re-running the curl line is safe. The installer reads ~/.openjarvis/.state/install-state.json and skips completed steps. If your venv got nuked, re-running heals it.

Cloud quick-path

If any of these env vars are set when you install or run jarvis init, the installer/init proposes cloud as the default and writes the matching provider into config.toml:

• OPENROUTER_API_KEY
• ANTHROPIC_API_KEY
• OPENAI_API_KEY
• GOOGLE_API_KEY (or GEMINI_API_KEY)

Local-first remains the default when no key is in env. Precedence is OpenRouter > Anthropic > OpenAI > Google.

Flags

Flag	Effect
--minimal	Skip the foreground model pull. First chat will need to wait for the bg pull to finish.
--no-bg-orchestrator	Don't detach the background work pipeline. (Mostly for testing.)
--force	Re-run all steps even if install-state.json says they're done.

Environment overrides

Variable	Default	Purpose
OPENJARVIS_HOME	$HOME/.openjarvis	Install location.
OPENJARVIS_REPO_URL	https://github.com/open-jarvis/OpenJarvis.git	Source repo for the clone step.

Uninstall

jarvis-uninstall

Removes ~/.openjarvis/, ~/.local/bin/jarvis, and ~/.local/bin/jarvis-uninstall. Leaves Ollama, uv, and the Rust toolchain in place (they may be used by other tools); the script prints removal hints.

Updating

jarvis update

Pulls the latest source, refreshes the editable install, and rebuilds the Rust extension in the background. Models are not touched.

Troubleshooting

"command not found: jarvis"

~/.local/bin isn't on your PATH. Run source ~/.bashrc (or ~/.zshrc) or open a new terminal.

"memory features unavailable"

Rust extension hasn't finished building yet (or failed). Check status:

jarvis doctor

Manually retry:

~/.openjarvis/.scripts/install-rust.sh && ~/.openjarvis/.scripts/build-extension.sh

A bigger model failed to download

Check status and retry:

jarvis doctor
~/.openjarvis/.scripts/pull-model.sh qwen3.5:9b

Behind a corporate proxy

Set HTTPS_PROXY and CURL_CA_BUNDLE in your environment before running the installer.