Ivy Tendril — The Agentic Software Factory for 10x Builders

AI agents can now write 99% of the code. This fundamental shift changes what it means to be a developer. Our role shifts to knowing what good looks like, specifying clear goals, and supervising execution. Ivy Tendril replaces your traditional IDE with an autonomous coding factory engineered for parallel, unattended agent execution.

Explore Documentation → Developer Portal →

Developer Portal & Fast Integrations

Whether configuring coding agents or integrating Tendril into your CI/CD pipelines, start with our developer resources:

• Developer Portal — Quickstart, self-serve keys, sandbox daemon, and rate limits.
• OpenAPI 3.1.0 Specification — Machine-readable REST API schema with typed RFC 9457 error models.
• Model Context Protocol (MCP) Manifest — Tools and Streamable HTTP endpoints.
• llms.txt — Machine-readable guidance optimized for LLMs.

Core Architecture & Key Capabilities

Parallel Git Worktrees

Tendril runs coding agents in isolated git worktrees. Keep your main working tree clean while agents implement features, fix issues, or explore refactoring in parallel. Changes are only integrated once you review, verify, and approve them.

Plan Lifecycle & Supervision

Every task is modeled as a structured Plan with distinct states: Draft, Approved, Running, Completed, Failed, and Icebox. Add annotations, inspect real-time agent output, and guide agent reasoning mid-execution.

Deterministic Verification Gates

Configure automated verification routines (compiler checks, unit test suites, linters, and Playwright visual screenshot tests) that run automatically after agent turns, providing immediate feedback before human sign-off.

Built-in Model Context Protocol (MCP) Server

Tendril includes a first-party Model Context Protocol server exposing plan management, worktree control, and agent execution tools to Claude, ChatGPT, and custom orchestrators.

Documentation Sections

• Getting Started — Installation, machine preparation, onboarding, and first plan walkthrough.
• Core Concepts — Plans, promptwares, and lifecycle management.
• Configuration — Project setup, danger zone management, and config schemas.
• Applications — Dashboard, Review, Plans, Jobs, Icebox, PRs, and Recommendations views.
• Coding Agents — Claude Code, OpenAI Codex, GitHub Copilot, OpenCode, and Google Gemini CLI integrations.
• Integrations — GitHub CLI, JamDev bug reporting, and OpenClaw capture.
• Model Providers — Berget, Evroc, Zai, Scaleway, Opper, OpenRouter, Cloudflare, NVIDIA, and Vercel.
• Advanced Reference — CLI commands, Axum REST & WebSocket APIs, and MCP server configuration.