> build_tools --target=teams_that_ship

Ship from theterminal, not the tab.

noir.dev is a deployment toolchain for product teams that prefer a single command to a forty-tab dashboard. Build, test, and promote to production without ever leaving the shell.

~/noir $ █

KESTREL.IO / LUMEN LABS / FERROUS / PALLAS / QUILLSTACK / ORBITAL / TIDEPOOL / VECTRA / BASALT / MERIDIAN.DEV / HALYARD /

> cat ./features.txt

Six primitives.One control plane.

Every feature is a first-class CLI command. Scriptable. Pipeable. Revertable. Nothing is locked behind a dashboard that a browser crash can lose.

// 01 / SYNC

Parallel git flow

Branch, review, and merge in sub-second steps. Push a change, preview a build, ship it, all inside a single command chain.

// 02 / PIPELINE

Tests as type-checks

First failure pins the branch before it ever reaches main. Retry on a slow flake, block on a real one, nothing in between.

// 03 / DEPLOY

Atomic releases

Promote a signed artifact with one command, roll back with another. No half-shipped builds, no drift between staging and prod.

// 04 / OBSERVE

Traces as timelines

Follow every request through every service without opening a dashboard. Pipe the trace straight into grep, awk, whatever you already trust.

// 05 / SECURE

Zero-trust default

SSO, scoped tokens, and audit logs enabled on every plan, free tier included. Nothing to bolt on when compliance comes asking.

// 06 / SCALE

Run anywhere

Bare metal, serverless, or the edge — one binary, one control plane, no lock-in. Bring your own cluster, keep your own data.

> man ./how_it_works

Three commands,from clone to prod.

No onboarding dashboard, no migration wizard. If your project builds locally, it deploys on noir.dev in three steps and roughly six minutes.

01

$ noir init

Clone the starter.

One command scaffolds a minimal project, writes your first lockfile, and wires the CI so the first commit is already green. No opinionated framework layer in the way.

02

$ noir build

Reproduce the build.

Every build runs in the same hermetic image as CI, so a green run on your laptop is a green run on main. Cache hits average 180ms on a warm host.

03

$ noir deploy --prod

Promote to prod.

A signed artifact ships to every edge region in parallel. The command returns only when health checks pass on every node, then prints the live URL.

> tail -f ~/demo.log

A live session,not a product tour.

This is the literal output of deploying this landing page to production — copied from the terminal, pasted below, nothing faked, nothing edited down.

$ noir init terminal-noir
  → scaffolding project structure          ok  (120ms)
  → resolving 64 packages                  ok  (1.8s)
  → wiring up .env and secrets             ok  (92ms)
  ✓ project ready — cd terminal-noir

$ noir build
  → compiling typescript                    ok  (640ms)
  → bundling edge functions                ok  (1.1s)
  → verifying signed artifact               ok  (84ms)
  ✓ build complete — artifact sha256:b7e1..4f2a

$ noir deploy --prod
  → uploading to 48 edge regions           ok  (3.2s)
  → running canary health checks            ok  (840ms)
  → warming edge caches                     ok  (1.1s)
  ✓ production → https://terminal-noir.noir.dev
  (total: 6m 12s)

UPTIME / 90 DAYS

Across every region, every minute. Measured, not marketed.
99.99%

P50 LATENCY

From the edge node to the first byte of response.
64ms

DEPLOYS / MONTH

Production ships across every customer, every region, averaged.
1.2M

EDGE REGIONS

Nodes on five continents, no preview regions in the headline.
48

> cat ./letters.txt

What the operatorswrote back.

We moved the whole deploy pipeline to noir in a weekend. The first thing I noticed was that I stopped opening the team chat to check if a shipment had failed — the command told me first.

Priya Menon Head of Platform · Fieldnote

Three months in we're averaging eleven production shipments a day and have not filed a rollback ticket since week two. I used to spend Mondays watching dashboards. Now I read them.

Daniel Okafor Staff Engineer · Riothouse

> pricing --plain-english

Three plans.Flat seats. No tricks.

Billed per active developer, per month. Annual billing cuts 20 percent. Every plan includes the full CLI, unlimited environments, and SSO.

monthly
yearly -20%

solo

$ 0 /dev/mo
  • 1 developer seat
  • Unlimited environments
  • CLI + local builds
  • Community support
  • SSO (Google, GitHub)
  • Custom roles
  • Audit logs
  • Dedicated SRE
most popular

team

$ 19 /dev/mo
  • Up to 50 developer seats
  • Unlimited environments
  • CLI + local builds
  • Priority chat support
  • SSO + SCIM
  • Custom roles
  • Audit logs (90-day)
  • Dedicated SRE

scale

$ 38 /dev/mo
  • Unlimited developer seats
  • Unlimited environments
  • CLI + local builds
  • 24/7 pager support
  • SSO + SCIM + SAML
  • Custom roles
  • Audit logs (unlimited)
  • Dedicated SRE