Maple Labs

Research

Where products start.

Mirrored Maple’s products begin as Maple Labs experiments. These are the problems we are working on, and where each one has gone.

Research areas

Four problems we keep coming back to.

Each one links to the Mirrored Maple product it became, or is becoming.

Detection that explains itself

An alert nobody can explain gets ignored. Every sentence our detection writes says which fields of which events it came from, and a sentence the events cannot support is not written. Deterministic, offline, no model required.

Ships as Sapience ↗

Credentials nobody else can read

A password manager whose server holds ciphertext only. Keys are made and kept on your devices, a new device joins only with an approval from one you trust, and recovery is a kit you hold.

Becoming MaplePass ↗

Endpoint telemetry

An agent that signs every request with its own key and collects what the operating system already logs, within a budget, so a busy log never loses events to a rate limit.

Ships as Seedling ↗

Offensive tooling

Penetration testing produces raw output; clients need findings. Tooling that keeps evidence, severity and the fix together, from first capture to the final report.

Ships as XPLT ↗

Previews

How an experiment reaches you.

  • It says it is a previewClearly labelled, everywhere it appears.
  • Support says so tooPreview support expectations are written down, and they are not production’s.
  • Telemetry is opt-inA preview collects nothing about its use unless you turn that on.
  • Rollback is safeTurning a preview off puts you back where you were.
  • No silent promotionA preview never quietly becomes production.
  • Retired with noticeWhen an experiment ends, its deprecation is documented first.

Questions about our research?

Write to the lab. For a vulnerability report, read the disclosure policy first.

hello@maplelabs.us Disclosure policy