II Agent

II Agent is an open-source AI agent built by Intelligent Internet and released under the Apache 2.0 license. It supports solo developers, research teams, and enterprises building internal tooling, with no vendor lock-in and full bring-your-own-API-key (BYOK) support. The project exited beta and is now generally available, with source code hosted on GitHub.

What It Is

II Agent is a general-purpose agentic AI framework that handles multi-step task planning and execution across a wide range of domains — from building mobile and web applications from a prompt, to conducting deep research, to manipulating documents and generating slide decks. It runs locally via Docker and supports multiple LLM providers including OpenAI, Anthropic, and Google, configurable through a YAML file or environment variables.

Core Capabilities

The README organizes II Agent's capabilities into four main areas:

• Build: Mobile app development, website app development, storybook generation, video and image generation, live editing for websites and slides, and visual plan mode before building.
• Research: Fast research and deep research modes, plus interactive website generation that turns research briefs into complete sites with citations and embedded Q&A.
• Automate & Integrate: Built-in and custom skills, GitHub-based workflow reuse, and app integrations including Gmail, Slack, GitHub, Notion, Google Calendar, Discord, Dropbox, and Canva.
• Everything Else: Multi-model chat with mid-thread provider switching, file attachments, code interpreter, PDF extraction and creation, Excel formulas and charts, Word editing, PowerPoint manipulation, and prompt-to-deck slide creation.

Architecture and Self-Hosting

II Agent is designed to run entirely on your own infrastructure. The stack includes a Python backend (port 8000), a Node.js/npm frontend (port 1420), PostgreSQL for persistence, Redis for task queuing, and MinIO for S3-compatible object storage. A Makefile provides commands for setup, development, migration, linting, formatting, testing, and full Docker Compose deployment. LLM providers are configured via model_configs.yaml or inline JSON in .env, with support for direct APIs, Vertex AI, Azure, and self-hosted models.

Update: v0.4

The repository's latest release is v0.4, published in July 2025. The project was created in April 2025 and has accumulated over 3,300 GitHub stars and 500+ forks as of mid-2026, according to the repository metadata. The README notes "significantly improved speed compared to earlier beta iterations" and describes the project as now out of beta. The GitHub topics tag the project under agent, agentic-ai, ai-agent-framework, llm, and manus.

Open-Source Deployment Model

Because II Agent is fully open source under Apache 2.0, users can fork, modify, and redistribute it freely. There is no hosted SaaS requirement — the web app at agent.ii.inc is available as a convenience, but the entire stack can be self-hosted. BYOK means users supply their own API keys for whichever LLM providers they choose, giving full control over cost and model selection.