English | 中文

Holon is a local workbench for agents doing continuous work.

Holon itself is not an agent. It provides a local working environment for multiple agents. Agents understand goals and drive execution; Holon treats "work" as the core unit, preserving state, organizing context, recording waits and wakes, so tasks that span sessions, commands, human confirmation, or external events can resume at the right time and eventually deliver results back to the operator.

Keep agent work alive in your local workspace.

Holon provides two interaction modes: TUI (terminal) and Web GUI (browser).

brew tap holon-run/tap && brew install holon

Or download binaries from GitHub Releases.

holon onboard

This walks through provider credential setup interactively. You can also configure providers through the Web GUI Settings page after starting the daemon. See Configuration Reference and Web GUI guide for more.

holon daemon start

holon tui

Select an agent and start working. Agents keep running after you disconnect.

Open http://localhost:7878. Create an agent and work through a chat interface with built-in file browser, task tracking, and more.

For more: TUI guide · Web GUI guide · First agent

brew tap holon-run/tap
brew install holon
holon --help

You can also download prebuilt binaries for Linux amd64, macOS amd64, and macOS arm64 from GitHub Releases.

The examples below assume holon is installed on PATH.

Holon needs a model provider before it can run agents. The recommended path is:

• holon onboard — interactive CLI setup that guides you through provider credential configuration without echoing secrets.
• Web GUI Settings — after starting the daemon, open http://localhost:7878 and configure providers through the Settings page.

Holon supports common providers such as Anthropic, OpenAI, DeepSeek, OpenRouter, Qwen, GLM, Xiaomi, Kimi, and MiniMax. For advanced setup including credential profiles, custom providers, and Codex subscriptions, see Configuration Reference and Supported Models.

Holon breaks agent work into a few explicit runtime objects:

• Agent — long-lived local identity with its own queue, state, and working context.
• WorkItem — continuously advanceable goal with a plan, progress, blockers, wait conditions, and a completion report.
• Task — supervised asynchronous execution (command, background task, or child agent).
• WaitFor / wake — explicit declaration of waiting for a task result, external event, or operator input, and resuming when the condition is satisfied.
• Workspace / worktree — execute in local repositories and isolate coding tasks into managed worktrees.
• Origin / brief — preserves input origin and trust information while keeping execution traces separate from operator-visible delivery.

For more detailed explanations, see Concepts.

Holon is under active development. The current recommended release is v0.30.0.

The current project focus remains the Rust runtime: agent lifecycle, queues, WaitFor/wake, tasks, WorkItems, trust boundaries, local workspaces, and structured delivery.

• Website docs — install, getting started, concepts, guides, reference
• Documentation layers
• Architecture overview
• RFCs
• Implementation decisions
• Release process

The Rust binary embeds web GUI assets at compile time via rust-embed. Build the frontend first, then compile the binary:

make all
holon --help

Or step by step:

make web    # build web GUI (requires Node.js 24 LTS)
make build  # build Rust binary

Use Node.js 24 LTS for Web GUI development. Run the same full validation used by CI with make:

make ci

For a focused Web GUI check, including Vitest and the production build:

make web-ci

See make help for the full list of targets.

Run the benchmark harness:

cd benchmark
npm install
npm test

• GitHub Discussions
• GitHub Issues

This project is licensed under the Apache-2.0 license.