Search inside videos by asking in plain words

How it works

To connect Postgres PGVector Store and Grist, you don't need a developer: a ready-made scenario links them in minutes.

1. Starts when: When clicking ‘Test workflow’
2. Then: Get Latest Youtube Videos
3. Then: Ignore Already Seen
4. Check: For Each Video
5. Then: Vectorise Subworkflow
6. Then: Get Video Subtitles
7. Then: Chunk Subtitles
8. Then: Chunks to Items
9. Check: For Each Chunk
10. Then: Qdrant Vector Store
11. Then: Wait
12. Then: Default Data Loader
13. Then: Embeddings
14. Then: Text Splitter
15. Then: Clean Up Output
16. Then: Sort By Video ID
17. Then: Generate Template
18. Then: Answer Query
19. Then: Markdown
20. Then: Map Fields
21. Starts when: Respond to Webhook
22. Then: Extract Results
23. Check: For Each Group
24. Then: Combine Results
25. Check: Has Results?
26. If yes: Transcripts to Items
27. If no: Generate Empty Response
28. Starts when: Respond to Webhook3
29. Starts when: SEARCH API
30. Then: Incr Rate Limit
31. Check: 10req/min
32. If yes: Get Query
33. Then: Get Embeddings
34. Then: Qdrant Groups Search
35. Check: Has Results?1
36. If yes: Groups to Items1
37. If no: Generate Empty Response1
38. Starts when: Respond to Webhook4
39. If no: 429 Response
40. Starts when: Respond to Webhook2
41. Starts when: Vectorise Subworkflow1
42. Starts when: WEB UI
43. Then: Generate Webpage
44. Starts when: Render Page
45. Starts when: Schedule Trigger
46. Then: OpenAI Chat Model
47. Then: OpenAI Chat Model1

You can launch this Postgres PGVector Store + Grist integration in Scriptera: describe the task in plain words — the scenario is built, launched and monitored for you.