Show HN: GibRAM an in-memory ephemeral GraphRAG runtime for retrieval

Summary

GibRAM Graph in-Buffer Retrieval & Associative Memory Graph in-Buffer : Graph structure (entities + relationships) stored in RAM : Graph structure (entities + relationships) stored in RAM Retrieval : Query mechanism for retrieving relevant context in RAG workflows : Query mechanism for retrieving relevant context in RAG workflows Associative Memory: Traverse between associated nodes via relationships, all accessed from memory GibRAM is an in-memory knowledge graph server designed for retrieval augmented generation (RAG) workflows. It combines a lightweight graph store with vector search so that related pieces of information remain connected in memory. This makes it easier to retrieve related regulations, articles or other text when a query mentions specific subjects. Why GibRAM? In memory and Ephemeral: Data lives in RAM with a configurable time to live. It is meant for short lived analysis and exploration rather than persistent storage. Graph and Vectors Together: Stores named entities, relationships and document chunks alongside their embeddings in the same structure. Graph aware Retrieval: Supports traversal over entities and relations as well as semantic search, helping you pull in context that would be missed by vector similarity alone. Python SDK: Provides a GraphRAG style workflow for indexing documents and running queries with minimal code. Components such as chunker, extractor and embedder can be swapped out. Quick Start Install via Binary # Install via script curl -fsSL https://gibram.io/install.sh | sh # Run server gibram-server --insecure Server runs on port 6161 by default. Install via Docker # Run server docker run -p 6161:6161 gibramio/gibram:latest # With custom config docker-compose up -d Python SDK pip install gibram Basic Usage: from gibram import GibRAMIndexer # Initialize indexer indexer = GibRAMIndexer ( session_id = "my-project" , host = "localhost" , port = 6161 , llm_api_key = "sk-..." # or set OPENAI_API_KEY env ) # Index documents stats = ...