intelligencedev/manifold

What is it?

Key features

• 01Agent chat for assigning objectives to specialists, with optional visualizations in responses.
• 02Visual workflow editor where MCP tools become nodes and saved workflows can be reused as tools.
• 03Scheduled tasks that run agent specialists on intervals or at set times, with Matrix support.
• 04Project workspaces that isolate files and load project-specific skills.
• 05Support for MCP servers with per-tool enablement to manage context size.
• 06Image generation through OpenAI, Google, and a ComfyUI MCP client.

Use cases

• Running long, multi-step research or analysis tasks with a team of specialized AI agents.
• Scheduling recurring AI-driven reports or checks delivered to chat channels.
• Building reusable automated workflows that combine MCP tools and other saved workflows.
• Running a self-hosted LLM stack with llama.cpp or vLLM alongside commercial model APIs.

Quick start

1. 1

   Copy the example env file

   cp example.env .env
2. 2

   Copy the example config file

   cp config.yaml.example config.yaml
3. 3

   Edit .env to set OPENAI_API_KEY and WORKDIR

4. 4

   Start Manifold with Docker Compose, then open http://localhost:32180

   docker compose up -d manifold

Things to know

• !Manifold is experimental and the README warns against production use requiring strong stability.
• !You need an LLM API key or a reachable OpenAI-compatible endpoint to do anything useful.

In collections

Topics

FAQ

What is manifold?

An experimental Go platform for long-horizon workflow automation using teams of AI assistants, with a Vue web UI. Manifold is a self-hosted workspace for running multi-step objectives with teams of AI specialists. Users assign work through a chat interface, and specialists can work on complex tasks over extended periods across multiple turns.

Is manifold free to use? What license does it have?

manifold is open source under the MIT license. Check the license terms before using it in commercial products.

Is manifold still maintained?

Actively maintained. The last commit to manifold was on Oct 22, 2026.

What can I use manifold for?

Running long, multi-step research or analysis tasks with a team of specialized AI agents. Scheduling recurring AI-driven reports or checks delivered to chat channels. Building reusable automated workflows that combine MCP tools and other saved workflows. Running a self-hosted LLM stack with llama.cpp or vLLM alongside commercial model APIs.

How do I install manifold?

Copy the example env file: cp example.env .env Copy the example config file: cp config.yaml.example config.yaml Edit .env to set OPENAI_API_KEY and WORKDIR Start Manifold with Docker Compose, then open http://localhost:32180: docker compose up -d manifold

Where can I get manifold?

The source code and installation instructions are on GitHub: https://github.com/intelligencedev/manifold