Propose
Generate candidate lab protocols, code patches, proof routes, and experiment variants through structured exploration.
Automated research systems
Research loops you can audit.
Mutome is a local-first workbench for eval-driven research loops: propose candidates, run validators and benchmarks, keep what improves, and archive what fails.
Randomness in discovery.Determinism in verification.
Request access
Premise
Model confidence is not proof.
The engine separates discovery from promotion. A model can suggest a protocol, patch, proof strategy, or experiment; Mutome only moves it forward when the run leaves replayable evidence behind.
Loop
Propose. Verify. Archive.
Verify
Evaluate candidates with deterministic validators, frozen benchmarks, replay commands, and predeclared metrics.
Archive
Keep reports, stdout, artifacts, hashes, failures, counterexamples, and decisions as reusable research memory.
Evidence
No promotion without a gate.
heuristiccandidate idea, untrusted until measured
model_checkedtests, validators, or adapters reproduce the claim
solver_certifieda solver certifies the relevant subclaim
proof_checkedmachine-checkable proof artifacts exist
reviewedhuman review, attribution, and reproduction are complete
Release path
Engine now. Desktop next.
The current engine runs protocol search, protocol autotune, and command autotune. It can evolve lab protocols against frozen validators, run git-backed patch/eval/metric loops, and package campaigns with artifacts, hashes, reports, and candidate ledgers.
The desktop release turns those outputs into a native research cockpit: launch runs, watch score lift, inspect the current mutation, connect local providers, and decide whether to keep, harden, or discard a candidate.