Headron Software Sign in

Give agents a shared foundation.

Context for the builder, a plan for the agent, proof for the change. Three systems, one pipeline.

View the quickstart

  1. InstallAdd the free, open-source CLI.
  2. Wire a repositoryheadron enforce install adds the commit hook, and headron init adds the checks for every file type you use.
  3. Connect agentsPoint Claude Code, Codex, Cursor or any MCP client at Headron.
  4. Share with chatGive ChatGPT, Claude, Gemini or Grok one link. It works to the same rules.

What you get

9×Fewer tokens
Prompt tokens against reading the whole repository, in a recorded routing test.
700Mistake kinds refused
Known mistake patterns with a check that refuses them, in a recorded snapshot.
353Files mapped to rules
Each file is tied to the rule it answers for, so a change knows which checks it owes.
122Failure ledger
Documented failure patterns and their guards, in a recorded snapshot.
6Declared controls
Tools, sandbox, budget, approval, effects and audit. Your runtime enforces them.
13Rules you write
Pick a task profile, then set the paths, commands, network and side effects it allows.
388Declared tools
Tools the contract declares, in a recorded snapshot. Your runtime enforces them.
94Honest instruments
Each instrument states its evidence and its limits.

These measurements show what was caught, not that a task succeeded.

View measurements

Connect

MCP
Public docs, or the open-source server for project-specific checks. MCP setup
Chat
Give a compatible assistant the agent guide and ask it to report what it has verified.
CLI
Run the open-source checks in your own environment.
CI
Configure your pipeline to run project checks on pull requests.
Mind API
A key-backed service for planning and agent integration. See API reference
Your agents
Command line agents, ACP agents and hosted providers, with your own keys and models.

Connect with MCP

Point any MCP client at https://headron.tech/mcp for public documentation. With a Mind API key it also plans gates. It never inspects your repository.

For project-specific discovery, install the open-source, read-only stdio server.

MCP setup guide Sign in

For machine-readable discovery, see Agent guide.

Engine and Mind API

The engine proves the change. Mind API tells the agent what to prove.

Open-source engine

Free under MIT, in your terminal, hooks and CI. Inspects the worktree, runs the gates and keeps the receipt.

Mind API

Hosted planning over REST and MCP, $10 a month. One policy revision for every agent, with nothing to install.

Enhanced MCP

Claude Code, Cursor, Codex and OpenHands read their gates inside the tool loop.

Keys you hold

Created in your account and copied once. Headron does not store them. Revoke and replace anytime.

Buy Mind API Mind API reference

Memory

The failure ledger records documented failure patterns. Before an edit, an agent can ask what broke in that file and what fixed it.

The ledger describes failure patterns and their guards. A guard covers the cases it tests, not every possible recurrence.

Verdicts are saved by id, never by paraphrase. An id can be re-checked; a summary drifts.

How a change lands

Contract

Before work starts: the goal, the paths and effects allowed, the budget and one acceptance test. Your runtime enforces the controls it declares.

Checks

Each file type owes checks from format to runtime. Configured hooks and CI run them, and the record shows what actually ran.

Verdicts

  • PassRan and met its test.
  • FailRan and missed. Blocks the merge.
  • SkipNot run or tool missing. Not a pass.
  • Not runNever reached. Not a pass.

Receipts

One record per run: the policy, the input and each check with its verdict, bound to the starting commit. Events are chained, so an edit breaks every later one.

Self-checks

Caught by name
Each planted mistake must be refused by the check built for it.
Plans, then code
The model plans a change; fixed code applies it.
Stated limits
Every check says what it proves and what it does not.
Only tighter
Quality limits move one way. Loosening one is done in the open.

Read the full proof map

Limits

Headron proves which checks ran and what they found, for recorded events only. A check is only as good as its test, so keep reviewing what matters.

Questions

Does Headron run my agent?

No. Your agents run where they already do: Claude Code, Cursor, your terminal. Headron decides which checks a change owes, holds unverified commits and keeps the receipt.

Which agents does it work with?

Any agent or runtime: Claude Code, Codex, Cursor, Antigravity, OpenHands, Windsurf, Hermes or a plain terminal. Each connects over MCP or REST and keeps its own models and tools.

What happens when a check tool is missing?

Headron records the check as skipped, and a skipped check is never a pass. Incomplete verification cannot clear your configured gates.

Does it slow agents down?

Agents get only the rules, gates and lessons for the files they touch, not the whole repository. A recorded routing test measured about 9× fewer prompt tokens.

Is it open source?

The verification engine is free and open source under MIT: CLI, gates and receipt checks. Mind API is the hosted service, $10 a month. Team collaboration and HeadronOS are in development.

How are Mind API keys stored and regenerated?

Keys are customer-held and generated on demand. Headron does not store secret keys. Copy each secret to your password manager or environment. If a key is lost or compromised, revoke it and generate a replacement.

What is the subscription and cancellation policy?

Mind API is $10 a month with no long-term commitment. Cancel anytime in your account settings; access runs through the paid period.