RAG indexing

Prerequisites

Repo activation

LOOPOVER_REVIEW_RAG=true, repo in LOOPOVER_REVIEW_REPOS (or features.rag: true in private config). Gittensor is_registered is not required for self-host RAG when the allowlist covers the repo.

Vector backend

SQLite vectors by default, Qdrant with the qdrant profile, or Postgres/pgvector where configured.

Embedding provider

An OpenAI-compatible embeddings endpoint with a model whose dimension matches the vector collection.

Choosing a vector backend

SQLite vectors are the default and need no extra service — fine for a small instance or getting started. Qdrant (QDRANT_URL, --profile qdrant) is the preferred dedicated vector store for review context at scale. A third option, PGVECTOR_ENABLED=true, uses the Postgres pgvector table instead — only relevant if you're already running the postgres profile and want to avoid standing up a separate Qdrant service. Leave it false (the default) when QDRANT_URL is set; Qdrant remains preferred for RAG at scale.

Qdrant and Ollama example

LOOPOVER_REVIEW_RAG=true
LOOPOVER_REVIEW_REPOS=owner/repo
QDRANT_URL=http://qdrant:6333
QDRANT_DIM=768
AI_EMBED_BASE_URL=http://ollama:11434/v1
AI_EMBED_MODEL=nomic-embed-text:latest

docker compose --profile qdrant --profile ollama up -d
docker compose exec ollama ollama pull nomic-embed-text:latest

Use QDRANT_DIM=1024 for 1024-dimensional models such as bge-m3 or mxbai-embed-large. If a Qdrant collection already exists, recreate it before changing dimensions.

AI_EMBED_API_KEY is the bearer credential for AI_EMBED_BASE_URL, if that endpoint requires one — a local Ollama typically doesn't, but a hosted OpenAI-compatible embeddings endpoint usually does. Setting AI_EMBED_MODEL alone does nothing without AI_EMBED_BASE_URL also set; unset, embeddings use the same provider as the rest of the review chain.

Indexing

RAG needs an index before it can retrieve useful context. A cold or missing index degrades to no context; the review still runs.

curl -X POST http://localhost:8787/v1/internal/jobs/rag-index \
-H "authorization: Bearer $INTERNAL_JOB_TOKEN" \
-H "content-type: application/json" \
-d '{"repoFullName":"owner/repo"}'

Operational checks

• Boot logs should include selfhost_embed_provider when an embedding provider is configured.
• Qdrant mode should log selfhost_vectorize with backend qdrant.
• Empty RAG context usually means the repo is not indexed, the embed model is unavailable, or dimensions do not match.

Pair RAG with AI providers and optionally REES.