What problem does it solve?
Provide conversational access to PostHog analytics so non-technical and technical users can run queries, investigate users and errors, and manage dashboards and feature flags without manually navigating the PostHog UI or writing HogQL from scratch.
Core Features & Use Cases
- HogQL query execution and generation: turn natural language questions into HogQL, run queries, and refine results for event counts, property analysis, and user timelines.
- Dashboard and insight management: create, update, and add insights to dashboards; follow a create → test → iterate pattern and respect the permanent-delete semantics.
- Person lookup & investigation: search persons by email or distinct_id, inspect their event timeline and properties, and cross-reference with other systems for attribution.
- Funnels, errors, and feature flags: build ordered or unordered funnels (with regex workarounds for OR filters), list and drill into errors while filtering by $host, and create/manage feature flags and experiments under an enforced permission model.
- Use Case: Investigate a sudden drop in conversion by confirming project, filtering by $host, running a funnel query, checking identity chain issues, and recommending engineering or product actions.
Quick Start
Query the last 7 days of step_start events for members.nsls.org, return counts, and propose a funnel from signup to onboard.