AI Integrations
Your AI model arsenal — we help you pick the right one per task to maximize quality and minimize cost. Connect providers in under five minutes.
GitHub
Connect repos and let AI push code as Pull Requests automatically.
View guide →MCP Server
Plan projects from inside Claude Desktop, Cursor, or any MCP client.
View guide →Anthropic (Claude)
Use Claude for autonomous AI task execution with top-tier reasoning.
View guide →GitHub Copilot
Leverage GitHub Copilot for AI-assisted code generation in your tasks.
View guide →
DeepSeek
High-performance open-source model for cost-effective task execution.
View guide →
Kimi (Moonshot AI)
128k context window — ideal for large codebases and complex projects.
Coming SoonOpenAI Codex
OpenAI's code-specialised model for precise, reliable task execution.
View guide →Google Gemini
Google's multimodal AI for versatile code generation and reasoning.
View guide →Use the MCP server where you build
Lovable, Bolt, v0, Replit, Base44, and Windsurf can call the hosted server while you chat. Each guide covers that product's screens, the token, and a full planning session.
Lovable
Plan in the Lovable chat connector before it generates the app.
View guide →Bolt.new
Add a custom HTTP connector and turn it on per Bolt project.
View guide →v0
Give v0 a bearer token. The generated app cannot call the tools.
View guide →Replit
Add a personal MCP server and an Authorization header for Agent.
View guide →Base44
One account connection for AI chat. Builder plan or higher.
View guide →Windsurf
Remote URL in Devin Desktop, or the local npx server.
View guide →