MaulwurfDIG AGENT
/
EARLY ACCESS
MAULWURF / HOW IT WORKS

Anatomy of a dig.

Four strata between your question and the answer. Here's what happens in each one.

−0m
STRATUM 0 — ACCESS

Connect, read-only

Grant read-only access to your repository. Maulwurf clones nothing permanently and changes nothing — it builds an index of your git history, source files, and documentation, then keeps it fresh as your code evolves.

−40m
STRATUM I — GIT HISTORY

Excavate the past

Commits are the fossil record of intent. Maulwurf scans history for when code changed, who changed it, and what the commit messages said — because "why" usually lives in the past, not the present.

−90m
STRATUM II — SOURCE & DOCS

Read everything

Source files, READMEs, ADRs, inline comments. The agent retrieves the passages most relevant to your question across all of them — not just the file you'd have guessed.

  • Full-text and semantic retrieval over code and docs
  • Cross-file reasoning: follows imports, calls, and configs
−140m
STRATUM III — SYNTHESIS

Answer with evidence

The answer is drafted from the retrieved evidence — and every claim is cited to the exact commit and line. If the evidence doesn't support an answer, Maulwurf says so instead of inventing one.

−200m
BEDROCK — HONEST LIMITS

What a dig can't do

  • It can't know what was never written down. Tribal knowledge stays tribal until someone documents it.
  • It won't modify your code. Maulwurf reads; it never pushes, merges, or refactors.
  • Citations can be wrong the way maps can be wrong — always verify before acting on them.

Bring your gnarliest repo.

We'll dig through it live and show you what comes up.

Get early access