MaulwurfDIG AGENT
/
EARLY ACCESS
MAULWURF / FAQ

Asked often.

Straight answers to the questions every engineering team asks first.

Can Maulwurf modify my code?

No. It authenticates with read-only scopes only — it cannot push, merge, open pull requests, or change a single line. It's an answering machine, not a coding agent.

Which repositories and languages work?

GitHub first, during the beta. Most mainstream languages are covered; exotic or generated code gets honest "I can't reliably read this" responses instead of guesses.

What if the answer isn't in the code?

Then Maulwurf says so. Every answer is built from retrieved evidence with citations; when the evidence isn't there, the item is marked uncertain — never invented.

How accurate are the citations?

Citations point to the exact commit and line the answer came from, so you can verify in one click. Treat them like a map: usually right, always checkable, never blindly trusted.

Does it work on monorepos?

Yes — arguably where it shines. Cross-service questions ("where does the auth token refresh happen?") are exactly what the excavation pipeline is built for.

Who can see our code?

Only your team. Indexes are isolated per customer, never used to train models, and you can purge everything at any time. See the Security page for details.

How does the beta work?

You apply as a design partner, we connect your repositories together, and your team starts asking questions — free during the beta. Your feedback shapes what we build next.

Something else on your mind?

We answer everything ourselves. No bots, no ticket queues.

Ask us directly