OpenFable
Open-source RAG engine using tree-structured indexes
Details
- External ID
- 47689633
- Source
- HN
- Company
- —
- Product
- OpenFable
- Website domain
- github.com
- Launched
- April 8, 2026
- Cohort
- —
- Upvotes
- 5
- Upvotes percentile
- 0.11182519280205655
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:26 p.m.
- Updated at
- Sept. 7, 2026, 9:26 p.m.
Description
Hi HN, I built OpenFable, an open-source retrieval engine that implements the FABLE algorithm (https://arxiv.org/abs/2601.18116) for RAG pipelines. I'm using it in another project and thought that others might benefit. Most RAG systems chunk documents into flat segments and retrieve by vector similarity. This works for simple lookups but breaks when answers span multiple sections, when relevant content is buried in a subsection, or when you need to control how many tokens you're sending to an LLM. OpenFable takes a different approach: when you ingest a document, it uses an LLM to identify discourse boundaries (not fixed-size windows), then builds a hierarchical tree, root, sections, subsections, leaf chunks, with embeddings at every level. Retrieval combines two paths: 1. LLM-guided path: the LLM reasons about which documents and subtrees are relevant from summaries 2. Vector path: similarity search with structure-aware score propagation through the tree Results from both paths are fused, deduplicated, and trimmed to fit a token budget you specify. You get the most relevant chunks, in document order, within budget. From the FABLE paper: the algorithm matches full-context inference (517K tokens) using only 31K tokens, 94% reduction, while hitting 92% completeness vs. Gemini-2.5-Pro at 91% with the full document. Retrieval only; OpenFable returns ranked chunks, not generated answers. Bring your own LLM for generation. It runs as a Docker stack (FastAPI + PostgreSQL/pgvector) and exposes both a REST API and an MCP server, so LLM agents like Claude Desktop or Cursor can use it directly. Trade-offs I want to be upfront about: - Ingestion is expensive; every document requires multiple LLM calls for chunking and tree construction - Retrieval isn't sub-second, the LLM-guided paths add round-trips - No built-in auth; designed to sit behind a reverse proxy - v0.1.0 — works end to end but the roadmap includes async ingestion, document deletion, and metadata filtering Stack: Python 3.12, FastAPI, SQLAlchemy, pgvector, LiteLLM, fastMCP. Apache 2.0. Happy to answer questions about the algorithm, implementation choices, or benchmarks.
Enrichment
- Theme
- browser automation and scraping for AI
- Vertical
- Horizontal
- Function
- Search & retrieval
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Hobby / open-source project
- Normalized one-liner
- tree-structured retrieval augmented generation engine
- Manually corrected
- False
Could you build this?
Partial A basic prototype wrapping the published FABLE algorithm is manageable, but building an efficient, production-ready tree-structured RAG retrieval engine with dynamic branching and low latency requires complex indexing logic.
What it would actually take: Implementing OpenFable requires implementing the exact hierarchical indexing paper specification: document parsing into recursive semantic trees, embedding node representations, and implementing tree-traversal search algorithms (e.g., beam search or branch-and-bound) rather than simple flat kNN vector searches. It requires optimized vector operations, storage layers handling hierarchical graph-tree relationships (e.g., SQLite/PostgreSQL with recursive CTEs or graph databases), and robust chunk-relationship caching.
Discussion
No comments on this launch.
Competitors
Other products that read as similar to this one — 94 launches clear the similarity bar, closest 8 shown.
Attention rank: #81 of 95 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 159 days after the earliest competitor.
- Breathe-Memory · hn · 2026-03-26 · 6 upvotes · similarity 0.51
- GibRAM an in-memory ephemeral GraphRAG runtime for retrieval · hn · 2026-01-18 · 60 upvotes · similarity 0.47
- A file-based agent memory framework that works like skill · hn · 2026-01-06 · 11 upvotes · similarity 0.45
- Smelt · hn · 2026-03-07 · 6 upvotes · similarity 0.43
- The biggest achievement of my life so far · hn · 2026-02-08 · 9 upvotes · similarity 0.42
- Librarian · hn · 2026-02-26 · 8 upvotes · similarity 0.42
- Project AELLA · hn · 2025-11-11 · 6 upvotes · similarity 0.42
- Extrai · hn · 2025-11-03 · 5 upvotes · similarity 0.42
Other launches for this product
- No other launches for this product.