Platform Overview

This page summarises RAG Loom's capabilities, primary components, and common deployment scenarios. Use it as a high-level orientation before diving into the detailed guides.

Core Use Cases

Scenario	Description	Where to Learn More
Knowledge base search	Ingest private documents and expose semantic search and summarisation	Quick Start, REST API
Production Q&A service	Serve retrieval-augmented responses with observability and scaling	Scaling, Security
Local experimentation	Run entirely offline with Ollama-backed models	Ollama Integration
Custom workflow integration	Embed the service within existing platforms or pipelines	REST API, API Script Playbook

Feature Highlights

• Modular pipelines for ingestion, retrieval, and generation, built on FastAPI.
• Pluggable vector stores (Chroma, Qdrant, Redis) and embedding models.
• Provider abstraction across Ollama, OpenAI, Cohere, and Hugging Face.
• Operational tooling including Docker Compose stacks, monitoring, and helper scripts.
• Extensibility through clearly defined service interfaces and configuration options.

Architecture at a Glance

RAG Loom orchestrates three primary flows:

1. Ingestion — Parse documents, generate embeddings, and persist to the configured vector store.
2. Retrieval — Perform top-K semantic similarity search over indexed chunks.
3. Generation — Assemble prompts from retrieved context and delegate to the chosen LLM provider.

See System Design Overview for a detailed diagram and component responsibilities.

Operational Building Blocks

Area	Summary
Deployment	Docker Compose stack with FastAPI, vector stores, optional Ollama, and observability tooling
Observability	Prometheus metrics, Grafana dashboards, and health checks
Scaling	Horizontal worker scaling, vector store tuning, model selection guidance
Security	Authentication hooks, network hardening, and secrets management

Links to the relevant runbooks are available in the Operations section.

Development Workflow

1. Prepare your environment (Prerequisites).
2. Launch locally (Quick Start).
3. Execute automated tests (pytest).
4. Review API schemas at /docs and explore example requests in the API Script Playbook.
5. Plan production rollout with the Scaling and Security guides.

Extending the Platform

• Implement new providers by extending the LLM adapter interface in app/services.
• Add custom business logic via FastAPI routers under app/api.
• Contribute documentation updates by editing the Markdown files in docs/docs and using ./utilscripts/docs_start.sh for live previews.

With the fundamentals in hand, choose the guide that matches your immediate goals—whether that's onboarding, operations, or integration.