HomeChatModelsImagesPackagesDocsInstallSettings

Install CanAI

One command to install 21 AI tools. Zero mandatory dependencies.

Quick Install

Works on macOS, Linux, and Windows (WSL). Python 3.10+ required.

Terminal
pip install canai

That's it. All 21 tools are now available as CLI commands. No extra dependencies needed.

Verify Installation

Run these commands to confirm everything is set up correctly.

Check version
canai --version
List all tools
canai list

21 Tools, 3 Layers

Every tool solves a specific pain point in the AI developer workflow.

🧠 Context Layer (8 tools)

Prepare context for AI

  • ctxbuilder — Codebase → CLAUDE.md, llms.txt
  • promptforge — Prompt template engine
  • smartpack — Intelligent context packer
  • scaffoldai — Bootstrap AI infrastructure
  • ctxmerge — Monorepo context merger
  • specforge — Spec → structured task list
  • repomap — Semantic repo map + Mermaid
  • webscout — Web scraping + DuckDuckGo search

🔗 Agent Layer (6 tools)

Connect to AI models

  • apibridge — Universal LLM adapter
  • smartroute — Task → cheapest model
  • tokencost — Token + cost tracking
  • memkeeper — Cross-session memory
  • cancraft — Terminal pair-programming
  • browsergaze — Browser + AI vision analysis

✅ Quality Layer (6 tools)

Verify AI output

  • qualgate — Security + quality gate
  • diffdigest — Git diff → AI summary
  • perfscope — Profiler → bottleneck report
  • logparser — Logs → structured analysis
  • architect — Expert system (8 phases)
  • gitrepo — Git operations

Requirements

Minimal requirements — CanAI is designed to be lightweight.

Python3.10 or newer
pip21.0 or newer
OSLinux, macOS, Windows (WSL)
Disk~5 MB
OptionalPlaywright (for browsergaze)
API KeysAt least one AI provider key