foreman
Open-source TUI supervises headless Claude Code agents through a gated delivery pipeline
Foreman is a serious process-first orchestrator for Claude Code, with real budgets, hash-sealed approvals, and crash-safe state. It’s Linux/WSL2-only with real setup cost, so it’s not for the curious. If you want no-setup cloud, look elsewhere.
Verified 8d ago · liveness 62/100 · cite: rightaichoice.com/tools/foreman
- Developers using Claude Code who want structured, governed project workflows
- Teams implementing AI-assisted development with human oversight and audit trails
- Individual open-source maintainers seeking automated PR pipelines with gated reviews
- Engineering orgs piloting agentic software delivery in a controlled environment
- Users seeking a no-setup, cloud-hosted agent platform
- Teams relying on non-Anthropic AI models (only Claude Code supported)
- Beginner developers unfamiliar with terminal UIs or Git workflows
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Foreman if you're not on Linux/WSL2, if you need multi-model support or a cloud-hosted no-setup platform, or if you prefer a web UI over a terminal interface.
No monetary cost, but you need your own Claude API key and pay Anthropic for every token used by agents.
Foreman is free and MIT-licensed, so the only cost is your Anthropic API usage plus your time. Compared to cloud agent platforms like Cognition's Devin or GitHub Copilot Workspace, which charge per seat or per run, Foreman is a zero-license option for teams that already have Claude Code access. But you trade that for significant setup and maintenance effort, so it's ideal for budget-conscious, technical teams rather than non-technical stakeholders.
In short
foreman — Open-source TUI supervises headless Claude Code agents through a gated delivery pipeline. Best for Developers using Claude Code who want structured, governed project workflows, Teams implementing AI-assisted development with human oversight and audit trails, Individual open-source maintainers seeking automated PR pipelines with gated reviews. Free to use.
What people actually say about foreman — is it worth it?
We scanned public community sources for foreman on Jul 1, 2026 and could not establish that the discussion we found is about this tool rather than something else sharing its name. Our own analysis of that scan says the posts were off-subject. Rather than publish a sentiment score built on the wrong subject, we publish nothing here and re-run the scan.
Viability Score
How well maintained and how widely used is foreman? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: September 2026
How we score →Key Features
- Headless Claude CLI integration in stream-json mode
- Gated pipeline: plan → ADR/PRD → issues → TDD build → e2e
- Human-in-the-loop review gates at every design phase
- Hash-sealed approvals that auto-revert if docs change
- Per-run turn/cost/time budgets enforced by orchestrator
- Daily cost ceiling with hard stop
- Crash-safe state via human-readable files in repository, no database
- Worktree isolation for non-colliding parallel workers
- PreToolUse deny hook blocks agents from writing own verification
- Evals flywheel: retro clusters failures into gated skill patches
- Keyboard-driven Textual TUI with full keymap
- Demo mode with mocked agent backend, zero token cost
- YAML-based configuration
- CLI commands: foreman init, build, status, retro, bench
About foreman
Foreman is an open-source terminal UI that supervises headless Claude Code agents through a gated software-delivery pipeline, bringing human oversight to AI-assisted coding. Instead of trusting agents to merge on their own, Foreman boxes them into a defined workflow: plan → ADR/PRD → issues → TDD build → e2e, with human review gates at every design phase. It spawns your locally installed Claude CLI in headless stream-json mode, parses its events, and enforces per-run turn/cost/time budgets along with a daily cost ceiling. All state lives as human-readable files in the repo—no database—so a crash mid-build recovers exactly where you left off. This tool is built for developers and teams using Claude Code who want governance, auditability, and safety without giving up automation. Foreman’s guardrails are enforced by the orchestrator, not by the agent’s “good behaviour.” A PreToolUse deny hook stops workers from writing their own verification, and worktree isolation lets parallel agents run without colliding, each gated by a declared touches set. The evals flywheel labels every run by outcome, clusters recurring failures into skill/prompt patches, and requires those to pass foreman bench before landing. The keyboard-driven Textual TUI gives you full control—launch, monitor, restart, or review any feature from the dashboard. A demo mode runs the entire pipeline on mocked agent output, so you can explore every gate and screen without spending a single token. Configuration is YAML-based, and CLI commands include foreman init, build, status, retro, and bench. Foreman is MIT-licensed and free, but it requires Python 3.11+, a locally installed and authenticated claude CLI, git, and Linux/WSL2. It is not a cloud-hosted or multi-model platform; it’s pointed at any repository to bring structure to agentic delivery. Compared to cloud agent platforms or generic AI code assistants, Foreman is for teams that value process discipline over raw speed.
Behind the Verdict
Foreman stands out because it shifts the trust boundary from the agent to the orchestrator. Instead of hoping Claude Code behaves, Foreman boxed it into a gated pipeline with enforced budgets and a pre-tool hook that blocks self-verification. This is the right approach for teams that need auditability and governance. Strengths: - Human-in-the-loop gates at design phases ensure architectural decisions are reviewed before code is written. - Hash-sealed approvals auto-revert if documents change, preventing silent drift. - Crash-safe state in plain files means no database to lose, and full recovery from disk. - Worktree isolation allows parallel agents without collision, gated by declared touches. - Evals flywheel turns every run into a learning opportunity, clustering failures into gated skill patches. - The demo mode lets you evaluate the entire pipeline without spending tokens. Weaknesses: - Linux/WSL2 only, which locks out macOS and native Windows users. - Only supports Claude Code; no other models or providers. - No formal API, limiting CI/CD integration. - YAML configuration has a learning curve. - TUI-only interface, no web or mobile. - Requires Python 3.11+ and a locally installed Claude CLI, adding setup complexity. Where it fits: Teams and individuals already invested in Claude Code who want to add structure, safety, and audit trails to agentic development. Perfect for piloting governed AI coding in a controlled environment. Where it doesn’t: If you want a cloud-hosted, no-setup platform that works across multiple models, or if you’re on macOS/Windows without WSL2, look elsewhere. For multi-model support, consider OpenHands or Aider, which are more flexible but less structured.
Researching foreman? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas foreman actually fits — and what changes day-one when you adopt it.
You have a backlog of issues but little time. You run 'foreman init', point it at your repo, and let it turn a high-level plan into ADRs, issues, and TDD stubs. You review each design gate via the TUI, approve with hash-sealed sign-off, and Foreman's agents implement features in isolated worktrees. You come back to a PR ready for review.
Outcome: You ship more features with less manual effort, while keeping full control over design decisions and a clean audit trail.
Your team adopts Claude Code but needs governance. You set up Foreman with per-run budgets and a daily cost ceiling. Agents plan and build in parallel worktrees, and every design phase requires human approval. failures are clustered via foreman retro into skill patches that must pass foreman bench.
Outcome: The team gets guardrailed autonomy, consistent process, and measurable safety without relying on agents' good behavior.
Use Cases
- Automate a gated software delivery pipeline from planning to e2e testing using Claude Code agents
- Generate ADRs and PRDs directly from a repository with human review
- Supervise AI agents performing TDD workflows with approval gates at each stage
- Onboard new projects by having Foreman create issues, ADRs, and test stubs from a high-level plan
- Maintain audit trails of AI-generated code changes through pipeline stage checkpoints
Models Under the Hood
as of 2026-08-31
Limitations
- Foreman is a TUI tool with no web or mobile interface.
- It depends on Claude Code's capabilities and Anthropic's API, with no support for other LLMs currently available.
- The project is in active development and lacks a formal API, which may limit CI/CD integration.
- Pipeline configuration is done via YAML, requiring familiarity with its schema.
- Linux/WSL2 only; macOS and native Windows not supported.
- Requires Python 3.11+, git, and a locally installed, authenticated claude CLI.
as of 2026-08-30
Verification history
We have re-verified foreman 7 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 7 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published foreman tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Open Source
$0
Ideal for
Developers and teams already using Claude Code and wanting a free, governed delivery pipeline without cloud dependencies
What this tier adds
Free and MIT-licensed; you only pay for Anthropic API usage. Includes all features: gated pipeline, budgets, worktree isolation, evals flywheel, and TUI—no paid tiers.
Where the pricing makes sense
The company stage and team size where foreman's pricing actually pencils out — and where peers do it cheaper.
Foreman is free and MIT-licensed, so the only cost is your Anthropic API usage plus your time. Compared to cloud agent platforms like Cognition's Devin or GitHub Copilot Workspace, which charge per seat or per run, Foreman is a zero-license option for teams that already have Claude Code access. But you trade that for significant setup and maintenance effort, so it's ideal for budget-conscious, technical teams rather than non-technical stakeholders.
Setup time & first value
How long it actually takes to get something useful out of foreman — broken out by persona, not the marketing-page minute.
On Linux/WSL2 with Python 3.11+ and git already installed, you can install Foreman and run 'foreman init' in under 5 minutes. Configuring YAML for your project's workflow takes 15-30 minutes if you want custom stages. Running 'foreman demo' lets you explore the full pipeline immediately without any setup—zero tokens spent.
Switching to or from foreman
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From a manual Claude Code workflow: run 'foreman init' on your repo to wrap existing agents in the gated pipeline
- →From custom scripts that call claude CLI: adopt Foreman's stream-json parsing and budgets to gain guardrails
- →From a cloud agent platform: export your process definitions and encode them as Foreman YAML stages
- ↗To a cloud agent platform: migrate staged approvals to their review features, but lose Foreman's file-based audit trail
- ↗To a different model provider: you'll need to switch orchestrators as Foreman only supports Claude Code
Resources & Guides
- Resourcegithub.com
README · foreman
Helpful link from github.com
- Documentationgithub.com
Docs · foreman
Full product docs from github.com
- Resourcegithub.com
CONTRIBUTING · foreman
Helpful link from github.com
- Resourcegithub.com
DECISIONS · foreman
Helpful link from github.com
- Resourcegithub.com
CHANGELOG · foreman
Helpful link from github.com
Tutorials & Learning
Official links
Tools that pair well with foreman
Common stack mates teams adopt alongside foreman, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Foreman vs Cognition Ai
If you are a developer deeply invested in Claude Code and want a free, open-source pipeline that keeps you in control with gated human reviews, Foreman is the way. For enterprise teams that need an autonomous engineer that plans, codes, tests, and ships across multiple platforms with a financial guarantee, Cognition AI's Devin is the clear winner despite the higher cost.
Foreman vs Poolside Ai
If you're an individual developer or small team already using Claude Code and want structured workflow governance, Foreman is free and powerful. For regulated enterprises needing custom on‑prem models, multi‑agent orchestration, and 256K context, Poolside AI is the right choice despite higher cost and vendor engagement.
Foreman vs Bito
Choose Foreman if you need a free, process-governed pipeline for Claude Code on a single repo with human-in-the-loop gates. Choose Bito if you work on multi-repo projects, need system-wide context for agents, and require enterprise-grade compliance or on-prem deployment.
Alternatives to foreman
View all
Frequently Asked Questions
Used foreman? Help shape our editorial sentiment research.