ForgeKit
Release tooling for humans
Public beta · works with GitHub Actions + GitLab CI

From commit to production—without the circus.

ForgeKit is a developer-first release toolkit: local commands, CI templates, and guardrails that make shipping boring (in the best way).

Used by Atlas Design Systems Monorepo Labs Shipwright.dev Neon Pine Games
forgekit · release preview

$ forgekit init

✓ Detected monorepo (pnpm workspaces)

✓ Installed policies: tests · lint · approvals

✓ Generated: .forgekit/policy.yml

$ forgekit release --preview

→ creating ephemeral env… ok

→ running checks… ok

→ posting to PR #184… ok

✓ Preview ready: https://pr-184.atlas.forgekit.run

Median
CI time: 7m
Policy
2 required checks
Rollback
1 command

No agents. No screenshots. Just your repo, your CI, and a set of guardrails you can read.

Tooling that respects your time.

Three layers: local DX, CI reliability, and release governance.

Browse docs

One-command releases

A CLI that speaks in verbs: init, preview, release, rollback. No wizards. No dashboards unless you want them.

example CLI
$ forgekit release
✓ policies passed
✓ release notes generated
✓ tag v1.18.0 pushed
✓ deployment triggered

Preview environments that post back to PRs

Every PR gets a URL, checks, and a clean teardown. Your review process becomes deterministic.

PR comments TTL teardown No vendor lock

Policy checks as code

Tests, lint, approvals, and “no deploy after 6pm” rules in a readable file.

Secrets sanity checks

Catch missing variables before the pipeline burns 12 minutes proving it.

Release notes from commits

Conventional commits in, clean changelogs out. Your customers get clarity.

How teams adopt ForgeKit

Start tiny. Earn trust. Scale the guardrails.

Opinionated by default

You can override anything, but the defaults are built from battle scars: flaky tests, leaking env vars, and release notes written at 2am.

  1. 1

    Install the CLI

    Bring ForgeKit to your laptop first. If it’s not pleasant locally, it won’t be trusted in CI.

    npm i -g forgekit
    forgekit --version
  2. 2

    Connect a repo

    Generate a minimal policy file and a CI template you can read in one sitting.

    forgekit init
    cat .forgekit/policy.yml
  3. 3

    Ship with checks

    ForgeKit blocks unsafe releases and explains exactly why—no mysterious red X.

    Deterministic output

    Same command. Same result. Less tribal knowledge.

    Roll back safely

    Rollback snapshots mean you don’t “fix forward” under pressure.

Pricing that scales with responsibility.

Free for experiments. Paid when you need guardrails and audits.

Monthly Annual (2 months free)

Solo

Free

For side projects and prototypes.

$0

  • 1 project
  • CLI + templates
  • Community support
Get started

Team

Most popular

For product teams shipping weekly.

/dev/mo

  • Unlimited projects
  • Preview environments
  • Policy checks + audit log
  • Priority support
Start Team plan

Tip: most teams see pipelines drop from ~18m → ~10m by week two.

Platform

Custom

For platform teams and compliance needs.

Let’s talk

  • SSO/SAML
  • Policy packs + approvals
  • Dedicated Slack channel
  • On-prem option
Contact sales

No surprise fees. Preview environments are metered by runtime, not “seats.”

See what’s included

Questions developers actually ask

Yes, we thought about your pipeline edge-cases.

Yes. ForgeKit ships readable templates for both, plus a minimal “bring your own runner” mode. If your CI can run a container, it can run ForgeKit.
Yes. We don’t require proprietary runners. Use hosted, self-hosted, or ephemeral runners—ForgeKit is just a CLI + config.
A simple YAML file defines required checks, approvals, and environment rules. The CLI evaluates it locally and in CI so there’s no “works on my pipeline.”
First-class. ForgeKit detects pnpm/yarn workspaces and lets you scope releases to packages, while still enforcing global policies.
Yes for Platform customers. The policy engine and audit log can run in your network, with your storage and identity provider.

Make shipping feel calm again.

Create a project, run one command, and get your first green release in under 10 minutes.

9m
median setup time
-44%
pipeline runtime
99.95%
service uptime

Stats from beta cohorts running 50–500 deploys/week. Your mileage may vary; your sanity should not.