Sweep embeddings × chunking × retrieval on your data — then ship the winning config first.
This is an external project guide. The runnable app lives in neomatrix369/rag-params-finder (MIT). This folder is the SIE-facing onboarding surface: short pages here, full detail in that repo.
SIE primitives used: encode (embeddings via BGE-M3 / Stella-v5 / SPLADE-v3);
optional score (SIE rerank). SIE is opt-in — the default stack runs without it.
| You are… | Start here |
|---|---|
| New to SIE, found this in the gallery | Getting started → SIE integration |
| New to rag-params-finder, want SIE embeddings | Same path — then What SIE does here |
Both audiences share the same local starting path (MongoDB stack, then the
dashboard). After that, optionally enable a remote SIE gateway and run one
example-sie.yaml sweep.
- Getting started — clone, prereqs, local Mongo path, dashboard
- SIE integration — env vars, health checks, first SIE sweep
- What SIE does here — models, encode/score, vs Voyage/local
- Troubleshooting — short FAQ + deep-links
Canonical docs in the project: QUICKSTART · SIE setup · docs index
| Service | Port | Notes |
|---|---|---|
| API | 8001 |
rag-params-finder server |
| Dashboard | 5374 |
Experiments UI |
| SIE (self-hosted only) | 8720 |
Not started by ./start-services.sh |
Built and maintained in neomatrix369/rag-params-finder under the MIT License. Screenshots and deeper guides live in that repository.