Enterprise Knowledge Retrieval
EDDI provides a complete Retrieval-Augmented Generation pipeline: knowledge bases that fill themselves from a website or from uploaded files, native support for multiple embedding providers and vector stores, and a zero-infrastructure RAG option via HTTP calls.
RAG Capabilities
- Web Crawler: Crawl a site within the limits you set, honouring robots.txt, its crawl delay and its sitemaps. Unchanged pages are skipped on the next run, and removed pages leave the knowledge base
- File Ingestion: Upload PDF, Word, Excel, PowerPoint, text, Markdown, CSV and HTML files, and each one is marked Indexed, Changed or Not indexed
- Scheduled Sources: Every source can run on a cron schedule, with a preview, a run history with counters and errors, and a purge
- 8 Embedding Providers: OpenAI, Azure OpenAI, Ollama, Mistral, Amazon Bedrock, Cohere, Google Gemini, Google Vertex AI
- 6 Vector Stores: pgvector, Chroma, In-Memory, MongoDB Atlas, Elasticsearch, Qdrant
- httpCall RAG: Zero-infrastructure RAG via any search API (BM25, Elasticsearch, custom endpoints)
- REST Ingestion API: Push documents yourself when a pull source does not fit
Flexible Deployment
RAG is fully configuration-driven. Choose your embedding provider and vector store via JSON configuration, no code changes needed. The httpCall RAG option lets you use any existing search infrastructure (Elasticsearch, Solr, custom APIs) without deploying a separate vector database.