Our take
Quietno commits in the last 90 days and no release in a year- Activity
- 20
- Community
- 19
- Issues
- 21
- Pull requests
- 21
+0 stars in the last 10 days
- Fewer contributors than 86% of the projects we track
goon has a broad, self-hosted workflow concept, but its quiet recent activity and highly concentrated contributions make it a cautious choice for production-critical automation.
Good fit if
- You want a Go daemon that can combine coding tasks and custom automations, with human approval before sensitive actions.
- Your team is comfortable configuring LLM and service credentials, and can run a Go 1.21+ application.
- You can evaluate and maintain a small, MIT-licensed project whose development depends mostly on one contributor.
Look elsewhere if
- You need evidence of active maintenance: the project has had no commits in the last 90 days and no release in over a year.
- Your organization requires broader contributor ownership or recent issue and pull-request activity; the measured period shows none, and most contributions come from one person.
- You need a turnkey hosted service rather than a self-hosted daemon that you configure and operate.
All health signals
| Last commit | 2026-06-15 (4 months ago) |
|---|---|
| Commits, last 90 days | 0 |
| Releases, last 12 months | 0 |
| Contributors | 2 (top contributor: 90% of commits) |
| Issues closed, last 90 days | 0 |
| Pull requests merged, last 90 days | 0 |
| Project age | 6 months |
Checked on 2026-10-17 with the GitHub API.
Overview
goon is a self-hosted Go daemon for running autonomous AI workflows alongside a team. Its built-in software pipeline polls Jira or GitHub Issues, works on tasks, runs tests, and can open pull requests; the same workflow engine can also handle scheduled jobs such as email digests or log checks.
The project is aimed at people who want a persistent worker rather than a one-off assistant. It offers web, CLI, and Telegram interfaces, and pauses at approval gates before sensitive actions. It can use several LLM providers, but the README does not identify one model or AI framework as the underlying engine. See the repository and project website for more detail.
Key Features
- Runs continuously as a daemon, or handles one-off tasks from the CLI.
- Polls Jira or GitHub Issues and can plan, code, test, and open pull requests on supported git hosts.
- Defines custom workflows in JSON, with executor, analyst, reviewer, loop, and notification stages.
- Supports scheduled automations, including recurring digests and monitoring jobs.
- Provides approval and rejection gates through the web UI, terminal, or Telegram.
- Stores project context and task history, and can reflect on recent activity to update learned notes.
- Routes planning, coding, chat, and review roles to different supported LLM providers or models.
Use Cases
- A small engineering team can have goon pick up Jira or GitHub Issues and prepare tested changes for human review.
- A developer can run a one-off coding or repository task from the terminal, using dry-run mode before allowing changes.
- A team can schedule a daily digest or periodic log scan by defining a workflow and connecting the relevant services.
- Someone who needs mobile oversight can approve work or interact with the agent through Telegram.
What you need
Detected in the repository
- Go 1.21 or newer (from go.mod)
- Automated checks on GitHub Actions
License in plain words
MITpermissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the copyright and license notice
- Share your changes: no
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
Install with Go and start the web interface:
go install github.com/harisaginting/goon@latest
goon start --web=:8080
Open http://localhost:8080 to configure an LLM and board. See the README for configuration, workflow examples, and other interfaces.
Alternatives
- multica: Multica focuses on assigning issues to existing coding-agent CLIs and reviewing runs, rather than running configurable workflows for coding and other tasks.
- agent-orchestrator: Agent Orchestrator is a desktop workspace for supervising coding agents across projects, rather than a self-hosted daemon for automated workflows.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| goon | Go | MIT | 3 | Quiet |
| multica | Go | Other | 51.9k | Active |
| agent-orchestrator | Go | Apache-2.0 | 12.7k | Active |
Considerations
- The measured project status is Quiet: there have been no commits in the last 90 days and no release in over a year. The repository is not archived, but its recent maintenance signals are limited.
- The contributor base is small and concentrated, so ongoing development and review capacity may depend heavily on one person. No issues were closed and no pull requests were merged in the measured recent period.
- The project requires Go 1.21 or later and credentials for whichever LLM and integrations you configure. The README describes zero runtime dependencies; Docker is not required.
- CI is present, but the supplied project facts do not report tests. Teams should verify the checks and workflows they rely on before adopting them for sensitive automation.
- Approval gates reduce unattended changes, but the daemon can run tools, access configured services, and make repository changes. Review permissions and workflow definitions carefully.
Found this useful?
Share it with someone who would like goon.