The harness
for AI coding agents.

Alpha Harness sits between your AI coding agent and your repo. It stops destructive commands before they run, brings your rules and notes into each session, and keeps your code on your machine. For Claude Code, Codex, Gemini CLI, Cursor and more.

Stopped before it ran
Same rules in 13 tools
26 ms per check · offline
Illustration. The verdict is the real one, returned by the shipped guard on Apple M2 Max, 2026-10-17.

Works with the tools you already pay for

  • ClaudeClaude Code
  • Google GeminiGemini CLI
  • OpenAICodex
  • CursorCursor
  • Terminal toolAntigravity
  • Terminal toolHermes Agent
  • Terminal toolKimi Code
  • Terminal toolGitHub Copilot CLI
  • Terminal toolOpenCode
  • Terminal toolAider
  • Terminal toolDroid
  • Terminal toolAmp
  • Terminal toolPi

Pricing

Free to start. Every plan adds more.

Every guardrail is free, in every plan. Paying adds the laws, the machines and the team around them. The first 100 Pro customers and the first 100 teams keep their price for as long as they stay.

Free

Try it on your own projects

$0
free, forever
No card needed.
Download freemacOS · Windows · Linux

Includes

  • Every guardrail, in the 9 tools that can block
  • The protection laws, in all 13 supported tools
  • Memory that survives every session
  • One workspace, 1 machine
  • Your own AI key, or local models
  • Anonymous usage data, switch off any time

Everything that protects your work is free and always will be.

Founding access

Pro

For developers who ship every day

$29
per month
or $290/year — two months free
$290
per year
that is $24.17/month — two months free

Everything in Free, plus

  • Premium laws, delivered and kept current from our cloud
  • Unlimited workspaces
  • Up to 2 machines of your own
  • Your price never increases
  • A direct line to the founders

After the first 100: $39/month. Yours stays $29 for as long as you keep paying.

For teams

Teams

For small teams that build together

$79
per month, for 3 people
That is $26.33 a person.
$790
per year, for 3 people
That is $21.94 a person a month, two months free.

Everything in Pro, for 3 people, plus

  • The team law: one shared rulebook, no rule written twice
  • Your company’s rules and blocked commands, in every tool
  • Admins, invitations and a workspace record you can export
  • Up to 2 machines per person
  • One invoice, one renewal date
  • Your price never increases

After the first 100 teams: $99/month. Yours stays $79 for as long as you keep paying. Grows 3 people at a time. Many more? Talk to us.

Enterprise

For companies of any size

Let's talk
priced for your team
The rate drops as you grow.
Talk to usAn answer the same day

Everything in Teams, plus

  • Any number of people and teams
  • Company rules set once, enforced over every team
  • Laws written with you for your industry
  • Volume pricing and invoicing on your terms
  • A named person to call

Tell us how many people and which AI tools they use, or email support@alphaharness.ai.

Compare every plan
FeatureFreeProTeamsEnterprise
Protection
Every guardrail, in the 9 tools that can block✓✓✓✓
The protection laws, in every tool✓✓✓✓
Laws
Premium laws from our cloud—✓✓✓
The team law (one shared rulebook)——✓✓
Your company’s rules and blocked commands——✓✓
Laws written with you for your industry———✓
Memory and machines
Memory across every session✓✓✓✓
Workspaces1UnlimitedUnlimitedUnlimited
Machines122 per person3 per person, more on request
People113Any number
Team
Admins and invitations——✓✓
Workspace record with export——✓✓
Live hive mind across machines——NextNext
Billing and support
Monthly or yearly (two months free)—✓✓As agreed
Founding price, kept while you pay—First 100First 100 teams—
14-day money-back guarantee—✓✓✓
Card or PayPal—✓✓Invoice on your terms
SupportEmailFounders, directlyFounders, directlyA named person

14-day money-back guarantee on paid plans · cancel any time · refer someone, earn 30%

Why Alpha Harness

Four layers between your agent and your repo.

Each one runs on your machine, each one is tested, and each one tells you exactly what it did.

01

Guards

The command that wrecks a week never runs.

Everything your agent tries to do is checked on your machine first. The destructive ones are refused, with a reason the agent understands, so it takes the safe path instead.

  • Force-pushes, mass deletes, overwritten secrets and schema pushes wait until you say yes. Open mode only warns.
  • Normal work is untouched: status, diffs and tests pass.
  • Careful mode holds commits, pushes and deploys for a human.
  • Under 26 ms per check (p95, Apple M2 Max). Works offline.
what your agent tried
git push origin main --forceBlockedrewrites shared history
rm -rf ./buildBlockeddeletes with no way back
echo "API_KEY=x" > .envBlockedoverwrites your secrets
npx drizzle-kit pushBlockedchanges a live database
git statusAllowed
npm testAllowed
git diff --statAllowed
Real verdicts from the shipped guard · 2026-10-17
02

Memory

Your agents stop asking what you already decided.

What you settled last month comes back at the moment it matters, and only then. When nothing is relevant, nothing is added: no padding, no bill for noise.

  • Uses the notes you already keep: Obsidian, Logseq, a plain folder. No notes yet? It starts a vault for you.
  • Every project sees only its own notes.
  • Secrets are filtered out before anything is remembered.
  • Your notes and your code stay on your machine.
what your agent was handed
  • “why do payouts fail on Base?”2 of your notes recalledPayout retries: what we decidedBase RPC limits
  • “rename this variable”Nothing relevant. Nothing added.
Illustration. The note titles are examples.
03

Laws

One rulebook. Every tool follows it.

8 protection laws on every plan, 18 with Teams, and room for your own. Each tool gets them where it actually reads them, and every copy stays identical.

  • One install finds the tools you have.
  • Your own text in those files is kept, never rewritten.
  • A session that was already open picks the rules up on its next prompt.
  • Uninstall removes exactly what was added.
what each tool gets
ClaudeClaude CodeRules + full guards
Google GeminiGemini CLIRules + full guards
OpenAICodexRules + full guards
CursorCursorRules + live guards
Terminal toolAntigravityRules + live guards
Terminal toolHermes AgentRules + full guards
Terminal toolKimi CodeRules + full guards
Terminal toolGitHub Copilot CLIRules + live guards
Terminal toolOpenCodeRules + live guards
Terminal toolAiderRules
Terminal toolDroidRules
Terminal toolAmpRules
Terminal toolPiRules
The installer's own list. Limits are spelled out on the download page.
04

Doctor

It checks itself, and fixes what is safe to fix.

Protection that quietly stopped working is worse than none. The Doctor tests every layer end to end on a schedule, repairs what is safe to repair, and shows you what it did.

  • Repairs only touch its own files. Never your notes, never a tool's settings.
  • Every repair is tested again before it is reported as fixed.
  • Tells you when something else on the machine is getting in the way.
  • One click to run it now, one click to copy the report.
doctor · example run
  • OKGuardsinstalled and answering
  • OKRulesidentical in every tool
  • FIXEDMemorysearch index rebuilt
  • WARNMachineanother harness is also installed
Illustration of one run.

Measured, not promised

Fast enough to forget.
Strict enough to trust.

Every check runs on your machine, in milliseconds. No server in the loop, so nothing waits on a network.

26 ms
per guard check, p95
Apple M2 Max
3,948
automated tests
3,476 engine · 472 app
57
guard rules
built in, no setup
8
protection laws, free
18 with Teams, every tool
13
AI tools covered
one install
41 / 41
risky commands stopped
Careful, the default level

Measured 2026-10-17 on the current build, on an Apple M2 Max. What changed since: the changelog.

Trust

Your code, your machine, your audit trail.

Your machine

Guards, memory and the Doctor run locally and keep working offline. Your code and your notes are never uploaded.

Your files

Rules are plain files on your disk. What we add is clearly marked, and your own text is never rewritten.

Your audit trail

Every block is logged with its rule. The Privacy screen prints the exact bytes of anything sent, and every part of it can be switched off.

Your exit

Uninstall removes every file and hook it added, in every tool, and says what it did.

Questions

Before you install.

Does it slow my terminal down?

No. Each check is measured at 26 ms at the 95th percentile on an Apple M2 Max. You will not notice it.

Does my code leave my machine?

Never. Rules, notes and guards are files on your disk and work offline. Every plan sends anonymous usage data with no file names, prompts or commands in it, and you can switch any part of it off. Every switch is on the app's Privacy screen, next to the exact bytes.

Which AI coding tools does it support?

Claude Code, Gemini CLI, Codex, Cursor, Antigravity, Hermes Agent, Kimi Code, GitHub Copilot CLI, OpenCode, Aider, Droid, Amp, Pi. The download page lists exactly what each one gets, and what it cannot do.

How do I undo it?

In the app: Settings → Uninstall. One click, and every file goes back the way it was. Your notes are never deleted.

Install once. Every AI tool you use is covered.

Download free