frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: OpenFable – Open-source RAG engine using tree-structured indexes

https://github.com/alainbrown/openfable
2•alainbrown•6h ago
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.

Show HN: Orange Juice – Small UX improvements that make HN easier to read

http://oj-hn.com/
30•latchkey•1h ago•22 comments

Show HN: TUI-use: Let AI agents control interactive terminal programs

https://github.com/onesuper/tui-use
23•dreamsome•2h ago•21 comments

Show HN: Go-Bt: Minimalist Behavior Trees for Go

https://github.com/rvitorper/go-bt
41•rvitorper•5h ago•3 comments

Show HN: I built a navigation app that displays weather along the route

https://navimodo.com/
27•vkatluri•2d ago•17 comments

Show HN: Explore the Silk Roads through an interactive map

https://www.intofarlands.com/silk-roads-map
39•intofarlands•4h ago•6 comments

Show HN: I pipe free sports streams into Jellyfin – no ads, just HLS

https://github.com/pcruz1905/hls-restream-proxy
81•pruz•7h ago•24 comments

Show HN: We built a camera only robot vacuum for less than 300$ (Well almost)

https://indraneelpatil.github.io/blog/2026/robot-vacuum/
92•indraneelpatil•2d ago•40 comments

Show HN: BAREmail ʕ·ᴥ·ʔ – minimalist Gmail client for bad WiFi

https://github.com/matt-virgo/baremail
38•Virgo_matt•4h ago•30 comments

Show HN: Eclipse Linux Alpha - Void-based with custom init

https://github.com/TheCodeVerseHub/EclipseLinux
4•SinisterMage2•1h ago•2 comments

Show HN: 500k+ events/sec transformations for ClickHouse ingestion

https://github.com/glassflow/clickhouse-etl
4•super_ar•2h ago•0 comments

Show HN: I built a tool to bootstrap VLESS and REALITY over SSH (with rollback)

2•insany•1h ago•0 comments

Show HN: An interactive map of Tolkien's Middle-earth

https://middle-earth-interactive-map.web.app/
274•frasermarlow•22h ago•58 comments

Show HN: We fingerprinted 178 AI models' writing styles and similarity clusters

https://rival.tips/research/model-similarity
67•nuancedev•5h ago•20 comments

Show HN: Gemma 4 Multimodal Fine-Tuner for Apple Silicon

https://github.com/mattmireles/gemma-tuner-multimodal
219•MediaSquirrel•1d ago•27 comments

Show HN: OpenMix, open-source computational framework for formulation science

https://github.com/vijayvkrishnan/openmix
2•vijayvkrishnan•2h ago•0 comments

Show HN: Brutalist Concrete Laptop Stand (2024)

https://sam-burns.com/posts/concrete-laptop-stand/
770•sam-bee•1d ago•232 comments

Show HN: Open-Source AI That Builds Screens, Not Just Text

https://github.com/SimonSchubert/Kai
3•arschibald•3h ago•0 comments

Show HN: PII-hound – A fast, dependency-free PII scanner in Go

https://github.com/saddledata/pii-hound
3•dbuckman•4h ago•4 comments

Show HN: A cartographer's attempt to realistically map Tolkien's world

https://www.intofarlands.com/atlasofarda
160•intofarlands•1d ago•31 comments

Show HN: Voxcode: local speech to text and ripgrep = transcript and code context

https://github.com/jensneuse/voxcode
6•jensneuse•8h ago•2 comments

Show HN: Unicode Steganography

https://steganography.patrickvuscan.com
27•PatrickVuscan•1d ago•5 comments

Show HN: OpenFable – Open-source RAG engine using tree-structured indexes

https://github.com/alainbrown/openfable
2•alainbrown•6h ago•0 comments

Show HN: Pion/handoff – Move WebRTC out of browser and into Go

https://github.com/pion/handoff
97•Sean-Der•1d ago•17 comments

Show HN: Ghost Pepper – Local hold-to-talk speech-to-text for macOS

https://github.com/matthartman/ghost-pepper
462•MattHart88•1d ago•195 comments

Show HN: Anos – a hand-written ~100KiB microkernel for x86-64 and RISC-V

https://github.com/roscopeco/anos
112•noone_youknow•4d ago•31 comments

Show HN: Stop paying for Dropbox/Google Drive, use your own S3 bucket instead

https://locker.dev
244•Zm44•1d ago•200 comments

Show HN: Finalrun – Spec-driven testing using English and vision for mobile apps

https://github.com/final-run/finalrun-agent
26•ashish004•1d ago•12 comments

Show HN: Tusk for macOS and Gnome

https://shapemachine.xyz/tusk/
119•factorialboy•4d ago•46 comments

Show HN: Hippo, biologically inspired memory for AI agents

https://github.com/kitfunso/hippo-memory
127•kitfunso•1d ago•25 comments

Show HN: I built a tiny LLM to demystify how language models work

https://github.com/arman-bd/guppylm
899•armanified•2d ago•134 comments