About

The layer that checks whether work is real before agents act.

Automation and AI agents now act faster than teams can reconstruct what happened. A tool call returns success. A deploy reports green. An agent says the task is done. None of that is proof.

Reported is not observed. An automation claims five steps completed — Mindplane shows three were observed. A deploy reports green — Mindplane shows validation evidence is missing. That gap is the product.

Mindplane is the verification layer for that gap. Not a workflow tool that makes work happen, but the system that proves whether the work matched what was declared, what was observed, what was decided, and what evidence the team can trust.

The unit is the receipt

Most tools show you the final output. Mindplane records the operational history of how that output came to exist: what was claimed, what was observed, what was decided, and what changed. Each of those is a receipt. A receipt is a unit of proven truth. Not a log line. Not a dashboard screenshot. Not a vibe.

Given a correlation ID, you can follow what actually happened across the chain. When a run diverges from the declared path, the receipt shows where. When an agent claims it ran five steps and only three were observed, the receipt is where the difference lives.

What we believe

Observed beats inferred. A system should tell you whether a value was measured or guessed, and it should not serve fiction as fact.

We do not claim what we cannot prove. Where Mindplane detects tampering, it says detected, not prevented. Where state is inferred, it says inferred. The product that proves reported is not observed cannot get caught reporting what it did not observe.

Observe before enforce. Mindplane runs in observe-only first. It shows you where it would have stopped automation before it blocks a single thing, so teams can adopt it without fear of breaking what already works.

Built for the people who get paged

Mindplane is for SRE, platform, and operations teams, and for the AI and ML teams now running agents in production. The shared problem is the same: work is happening faster than anyone can confirm it was the right work, done correctly, with evidence that survives scrutiny.

Audit trails explain what happened. Mindplane receipts give the next action something proven to read. An audit trail is read by a reviewer after the fact. A receipt is read by the next move: the agent checking whether the work it is about to do is still needed, the engineer who has to know what is true before they touch anything.

Where we are

Mindplane is early, and it is real. V1 is a multi-tenant operational truth layer: receipts, correlation, and a declared-versus-observed diff, built around real ingested events rather than staged claims. The record is not write-only. The system already reads its own truth back: it can re-assemble a run from a single ID, recompute whether a past decision still holds, and verify the record has not been altered since it was written.

We are deliberate about the line between what is shipped and what is still ahead, and we would rather show you a smaller true thing than a larger claimed one. That is not modesty. It is the product. A truth layer that overclaims has already failed at its one job.

The team

Mindplane is built by Scott Alexander, Founder and CEO, an SRE leader with over ten years building scalable infrastructure under production pressure, and Jay Kinder, Co-Founder and CTO, a principal SRE focused on automation and platform engineering. Both founders are verifiable on LinkedIn through the links above.