Show HN: Local CPU model for fact-checking, summarizing, explaining text

Category: cli-tool

Tags: cli, llm, offline, privacy, text-processing, fact-checking, translation

Score: 6.5/10 (Innovation: 6, Technical: 6, Documentation: 8, Utility: 6)

fftext is a CLI tool that provides CPU-only text summarization, explanation, fact-checking, and translation using a quantized 0.8B language model via llama.cpp, requiring no GPU or cloud API. Its innovative fact-check feature extracts claims, web-searches them, and labels verdicts with citations entirely offline (except web search), making it a lightweight, privacy-focused alternative to larger LLM pipelines. The project is interesting for its practical, self-contained approach to common NLP tasks with minimal dependencies.

Target audience: backend devs, data engineers, power users

Repository: https://github.com/kouhxp/fftext · Python · Apache-2.0 · 7 stars

View on Hacker News