install

Install Squeezy and get started.

Choose your platform, install the native agent, run diagnostics, then start Squeezy inside the repository you want it to understand.

choose platform

Release targets and primary install commands.

curl installer

macOS

Apple Silicon, Intel

curl -fsSL https://raw.githubusercontent.com/panicless/squeezy/main/install.sh | sh

aarch64-apple-darwinx86_64-apple-darwin

Or: brew install panicless/tap/squeezy

curl installer

Linux

x86_64, ARM64 static binary

curl -fsSL https://raw.githubusercontent.com/panicless/squeezy/main/install.sh | sh

x86_64-unknown-linux-muslaarch64-unknown-linux-musl

Or: cargo install squeezy --locked

Winget

Windows

x86_64

winget install panicless.Squeezy

x86_64-pc-windows-msvc

Or: extract squeezy-x86_64-pc-windows-msvc.zip and add to PATH.

first run

Verify the install before you start working.

squeezy doctor checks configuration, repo profile, provider credentials, session storage, and sandbox availability. After that, run squeezy from the repository root.

start

# One key, many hosted models via OpenRouter
export OPENROUTER_API_KEY=...
export SQUEEZY_PROVIDER=openrouter
squeezy config init --user
squeezy doctor
squeezy

providers

Bring your model key

Start with an aggregator (OpenRouter, Vercel AI Gateway, PortKey) or use first-party vendor APIs (OpenAI, Anthropic, Google Gemini), cloud-platform hosts (Azure OpenAI, Amazon Bedrock), local Ollama, or any OpenAI-compatible service (Groq, xAI, DeepSeek, Mistral, Together, Fireworks, Cerebras).

ask squeezy

Use local help

/help and natural questions about Squeezy are answered locally from bundled docs and your redacted config when they match a known topic, with a small doc-help pass for anything they don't cover.

report

Preview before sending

Use /feedback, /report, or squeezy sessions report to preview redacted support material before anything is uploaded.