Install SQ2.ai

Install the SQ2 execution system locally and run it on your repository in minutes.

Requirements

OS

macOS or Linux

Python

3.10 or higher

Git

Any recent version

GPU

Optional. CPU-only works.

No GPU required. NVIDIA GPU improves performance but SQ2 runs fully on CPU.
01 — Install

One command

Installation is local. No external execution required.

bash
curl -fsSL https://raw.githubusercontent.com/ft-dev-org/sq2-ai/main/scripts/install.sh | bash
02 — Verify

Confirm installation

bash
sq2 version SQ2.ai v1.0.0
03 — Initialize

Point SQ2 at your repository

bash
cd your-repo sq2 init . Initializing SQ2.ai for /home/user/your-repo... Scan complete: 142 files found. 142 files indexed, 580 chunks. Ready.
04 — System check

Verify environment

bash
sq2 doctor Python 3.12.7 OK Ollama 0.18.2 OK Ollama API running OK GPU RTX 4090 OK App dir ~/.sq2 OK Git 2.43.0 OK All checks passed.
05 — Scan

Inspect project files

bash
sq2 scan .

06 — Run workflows

Execute on your codebase

Execution is scoped, explicit, and verified. Verification is required before completion.

Analysis
sq2 run --workflow analysis

Analyzes project structure, dependencies, and code quality. Read-only.

Tests
sq2 run --workflow tests

Generates missing tests, executes them, and verifies results.

Fix
sq2 run --workflow fix

Identifies failing tests, generates fixes, and re-verifies until all pass.

07 — Safe execution

Verify before executing

Add --dry-run to any workflow to see the full execution plan before any files are modified.

bash
sq2 run --workflow analysis --dry-run
08 — Direct tasks

Execute specific work

bash
sq2 task "add tests for parser module"

Plans the task, executes scoped changes, and verifies results before completion.


Troubleshooting
sq2: command not found
Activate the virtual environment: source .venv/bin/activate
Python version too old
SQ2 requires Python 3.10+. Check with python3 --version
Permission denied
Run the install script without sudo. SQ2 installs into a local virtual environment.
Ollama not found
Install Ollama from ollama.com, then pull a model: ollama pull qwen2.5:14b

Run SQ2 on your repository

From install to execution in under 2 minutes.

REQUEST ACCESS
SQ2.ai © 2026