Compare
The proactive agent, with receipts. Databuddy's "Databunny" agent is squarely in our category: it investigates the data on its own and posts "what happened, why it matters, what to do" into Slack. It is our closest competitor and a good product.
| Capability | Databuddy | Vitrus |
|---|---|---|
| Proactive digest (arrives unasked) | ✓ | ✓ |
| Clickable query behind every number | — | ✓ |
| Unprovable AI sentences auto-dropped | — | ✓ |
| Local LLM / bring your own key | — | ✓ |
| AI referral ↔ AI crawler split | — | ✓ |
| MCP server for AI agents | ✓ | read-only, evidence attached |
| Self-hostable | AGPL, Postgres+ClickHouse+Redis | Apache-2.0, embedded |
| Product analytics depth | broader | narrow |
Features we don't have are shown as “—”. A comparison that hides its own gaps would contradict the one thing this product claims.
Their agent goes deeper, and their product analytics — funnels, errors, vitals — is broader than ours. They also have wider ecosystem integrations.
When you want to verify what the agent said. Databunny's output is raw LLM generation; ours is bound to an evidence bundle and anything unverifiable is dropped before it ships. You can also keep the data and the model on your own server.
Both are proactive. The difference is where the number comes from: Databunny lets the model read the data and write the sentence; we run fixed queries first and the model only rephrases that bundle. On the way out, every number is checked against the evidence.
Other comparisons: