Mneme HQ / Pilot
A Mneme pilot starts with work your team has already done. Share a repository or attach an architecture audit. We review it before a short follow-up, agree which recommendations matter most, and turn 3–5 of them into controls your team can test safely before anything blocks delivery.
Early access

Request a Mneme Pilot

Send the request and we take it from there. No presentation, rewritten requirements, or second audit is needed.

What your pilot delivers

  1. Stage 1

    Establish the baseline

    We audit a real repository of yours and identify the architectural decisions worth protecting.

  2. Stage 2

    Operationalize the decisions

    Applicable decisions become enforceable Mneme controls, connected to the engineering workflows your team already uses.

  3. Stage 3

    Produce evidence

    We test representative changes and demonstrate which architectural constraints are now being protected—in observe mode first, before anything blocks delivery.

Good-fit pilots usually involve
  • Cursor, Claude Code, GitHub Copilot, or agentic coding workflows in active use
  • Existing ADRs, repo rules, or engineering standards to enforce
  • Architectural drift, review delays, or consistency issues appearing in AI-generated code
  • Platform, backend, infrastructure, or data-heavy systems with non-trivial architectural constraints
A short 20-minute follow-up. We review your material first, so the call can focus on priorities and practical next steps. Complimentary for qualified teams; we respond within 1–2 business days.
Private repositories are welcome. Mneme’s deterministic enforcement runs in your development or CI environment. A repository does not need to be public to participate in a pilot.
A few details, then we take it from here
Your architecture audit is ready

We already have the repository, findings, and recommendations. You won’t need to enter them again.

A sentence is enough. If you arrived from an audit, the findings are already attached—edit the suggested text only if you want to.
Next: we review your audit before a short 20-minute follow-up, then agree which recommendations to address first.
Complimentary for qualified teams · response within 1–2 business days.

Frequently asked

Who is the Mneme pilot for?

Engineering teams already using AI coding tools (Cursor, Claude Code, GitHub Copilot, or agentic workflows) that have existing ADRs, repo rules, or engineering standards to enforce, and are seeing architectural drift, review delays, or consistency issues in AI-generated code. Strongest fit: platform, backend, infrastructure, or data-heavy systems with non-trivial architectural constraints.

Do I need a pilot to use Mneme?

No. Mneme is open source under the MIT license and self-hosted. You can install it yourself any time with pip install mneme-hq>=0.9.0 and run mneme check against your repo. The pilot is for teams who want help applying it across their repositories, ADRs, and agent workflows.

Which AI coding tools does Mneme work with?

Mneme is repo-native and model-independent. It works alongside Cursor, Claude Code, GitHub Copilot, and agentic IDEs, enforcing the same architectural decisions regardless of which tool or model generates the code.

What do you need from us to start?

If you came from an architecture audit, the repository, findings, and recommendations are already attached. Add a few team details and any priority you want us to consider. Otherwise, share a repository and one recurring architecture or review problem.

What happens after I submit the form?

We review the repository or attached audit before replying within 1–2 business days. A short 20-minute follow-up confirms the priorities. We then turn 3–5 recommendations into controls and test them in observe mode before anything blocks delivery.

An engineer writing on a glass wall while a colleague looks on
A pilot starts with one real boundary your team already agreed to protect.