Find the root cause, not just the symptom.
Waggle tests every suspected cause against your code and data, and reports only what survives an adversarial audit.
Open-source edition · runs on your machine with your own keys.
From dozens of suspicions to the few that hold up.
Stage 1 of 5: Comprehend. Maps your code, data and system.
Maps your code, data and system.
Root causes across code and data
Tests each suspected cause against your repos and data (BigQuery, SQL, Redis or any MCP server), then audits every claim.
Keeps checking, never starts over
Checkpointed scans resume where they stopped. Reruns, by hand or from CI, reuse unchanged work and flag what's new, fixed or still open.
Memory in two layers
Editable, versioned memory cards plus cross-project facts that flag disagreements. Later scans re-check both, never treat them as proof.
A fix plan your coding agent can run
Each finding comes with evidence, a fix and a check, plus a REMEDIATION.md your coding agent can run.
