1. Ingest
GitHub and Slack connectors pull commits, PRs, issues, reviews, merges, and channel messages into one event store.
Open →How Trace works
Trace turns GitHub and Slack activity into workflows, a knowledge graph, and operational signals you can investigate.
From events to intelligence
2. Reconstruct workflows
Related events are linked into workflows with timestamps and evidence — e.g. issue → PR → review → merge → Slack deployment thread.
Open →3. Build the organizational graph
People, repos, tickets, channels, and workflows become nodes; performed, touches, and participates_in edges show who touched what.
Open →4. Analyze
Bottleneck, knowledge-risk, and process-mining engines run on workflows and graph structure to surface delays, concentration, and drift.
Open →5. Investigate
The Investigation Agent answers questions using the same evidence — with LLM synthesis when configured on your workspace.
Open →How we infer dependencies and blockers
Trace does not guess org charts from commit messages alone. It uses timestamps, actors, and links in your ingested data.
Who work depends on
From co-occurrence in workflows and graph edges: same PR reviewers, actors on shared repos, handoffs from GitHub events to Slack threads. Knowledge risk highlights actors/repos with disproportionate activity.
What blocks delivery
Bottlenecks measure time gaps between staged events (e.g. PR opened → reviewed → merged). Cross-tool delays (GitHub ↔ Slack) and repo/ticket hotspots show where work stalls — not guesses from commit messages alone.
Empty insights?
- Connect GitHub or Slack and run sync with the jobs worker.
- Refresh analytics on the dashboard after new events arrive.
- Insights need enough cross-tool workflows — not just isolated pushes.