First Collection Workflow
What this page is for
Run an end-to-end workflow: create collection, ingest content, review session, publish, search.
Prerequisites
- Backend running at
http://localhost:3000 - Redis and Qdrant healthy
X-User-IDvalue for API calls
Steps
- Create collection.
curl -X POST http://localhost:3000/api/collections \
-H 'Content-Type: application/json' \
-H 'X-User-ID: demo@ragler.ai' \
-d '{"name":"Support KB","description":"Support docs"}'
- Ingest manual source.
curl -X POST http://localhost:3000/api/ingest/manual \
-H 'Content-Type: application/json' \
-H 'X-User-ID: demo@ragler.ai' \
-d '{"content":"Redis stores draft sessions. Qdrant stores published chunks."}'
- Generate chunks if needed.
curl -X POST http://localhost:3000/api/session/<SESSION_ID>/chunks \
-H 'X-User-ID: demo@ragler.ai'
- Review and adjust chunks.
curl http://localhost:3000/api/session/<SESSION_ID> -H 'X-User-ID: demo@ragler.ai'
curl -X PATCH http://localhost:3000/api/session/<SESSION_ID>/chunks/<CHUNK_ID> \
-H 'Content-Type: application/json' \
-H 'X-User-ID: demo@ragler.ai' \
-d '{"text":"Updated chunk text"}'
- Preview and publish.
curl -X POST http://localhost:3000/api/session/<SESSION_ID>/preview -H 'X-User-ID: demo@ragler.ai'
curl -X POST http://localhost:3000/api/session/<SESSION_ID>/publish \
-H 'Content-Type: application/json' \
-H 'X-User-ID: demo@ragler.ai' \
-d '{"targetCollectionId":"<COLLECTION_ID>"}'
- Search published content.
curl -X POST http://localhost:3000/api/search \
-H 'Content-Type: application/json' \
-H 'X-User-ID: demo@ragler.ai' \
-d '{"query":"Where are draft sessions stored?","collectionId":"<COLLECTION_ID>","limit":5}'
Verify
- Publish response contains
publishedChunks> 0. - Search returns relevant results.
Troubleshooting
Invalid collection ID format: ensure UUID fortargetCollectionId.- Empty search results: confirm publish completed and query matches stored content.
Next steps
/docs/product/intro/docs/architecture/overview