Developer / Stakeholder Tool
Data & Retrieval Console
MongoDB ↔ Qdrant sync, org-data health, and a live look into the RAG pipeline.
API base URL
Session token (X-Basik-Session)
Save
Test
Sync & Health
RAG Visualizer
Re-ingest
Load the latest reconciliation.
↻ Refresh sync
Index quality · BM25 + chunk health (sampled)
Loading…
LLM queue health · throughput / latency / saturation
↻ Refresh queue
Live while this API process runs embeddings (i.e. during a re-ingest).
Load now
Per-mailbox coverage
✨ Visualize
Runs planning + retrieval only (no synthesis), so it's fast and shows exactly what was understood and retrieved.
1 · Query understanding
2 · Blend & rerank
3 · Pipeline steps
Retrieved chunks
Trigger re-ingestion (production only)
Action
Sync — make Qdrant match MongoDB (recommended)
Dedup attachments — collapse duplicate files
Emails: full resync (force re-vectorize all)
Emails: retry missing chunks
Emails: fix degraded (empty-content) chunks
Attachments: re-vectorize
Meetings: re-vectorize extracted
clearExisting (wipe first)
Start job