Recipes
Self-contained, runnable scenarios using K3’s production Scriptum catalog. Every recipe is grounded in a real template (no inventions) and uses the same dodil k3 flow you saw in the Quickstart.
What’s here
| Recipe | Template used | Destination | When you need it |
|---|---|---|---|
| PDF → Vector | text_embedding_index | Vector collection | Auto-index PDFs / docx / HTML for semantic search (the canonical RAG ingest) |
| Documents → Warehouse | entity_pii_extraction | Warehouse table | Extract structured entities + PII into queryable rows |
| Replay & Retry | — | — | Recover from FAILED / PARTIAL ingest jobs, replay after pipeline changes |
Other templates worth knowing
The recipes here pick one template each as the worked example. For other production scenarios, the same flow works with any template from the catalog — swap the --scriptum argument and the destination kind:
| Goal | Template | Destination |
|---|---|---|
| Code search across a repo | code_embedding_index | Vector |
| Multimodal asset library | visual_embedding_index | Vector |
| Face recognition | face_embedding_index | Vector |
| Object detection on images | object_embedding_index or object_detection | Vector / Warehouse |
| Audio transcription | audio_transcription | Warehouse |
| Multi-level summarization | summarization | Warehouse |
| Document triage / routing | document_triage | Warehouse |
| Sentiment dashboards | sentiment_intent_analysis or review_analysis | Warehouse |
| Image understanding (OCR + detection + reasoning) | image_understanding | Warehouse |
| Translation | translation | Warehouse |
| Code intelligence (symbols, deps) | code_intelligence | Warehouse |
| Product catalog enrichment | product_catalog_enrichment | Warehouse |
| Video surveillance | video_surveillance | Warehouse |
Full descriptions, modalities, and labels: API Reference → Templates → The catalog.
Conventions used in these recipes
- CLI-first — every step has a
dodil k3 ...command. Anything not in the CLI today drops tocurlagainst the API. - Endpoint examples use staging
https://k3.dev.dodil.io. Production:https://k3.dodil.io. - Buckets use the placeholder
kb-prod; replace with your real bucket name. - Internal source is automatic — every bucket has one from
CreateBucket. You only need to look up itssource_idto wire rules. - External sources are Preview — these recipes use the internal source exclusively; the same pipeline + rule + job model works once external sources go GA.
Other useful pages
- Quickstart — first pipeline + first ingest in 5 minutes
- Core Concepts — Source, Credential, Pipeline, Template, Rule, IngestJob
- API Reference — Source / Pipeline / Ingest contracts
- CLI Guide — every
dodil k3command in this domain