Lumentis is an open-source command-line tool that turns meeting transcripts and unstructured text into comprehensive, easy-to-skim documentation with a single command. With about 1,700 stars on GitHub and updates into 2025, it uses large language models such as GPT-4o and Gemini Flash to do the heavy writing for you.
You run it with npx lumentis, point it at transcripts or large documents, and answer a few interactive prompts about the topics and structure you want. Lumentis then queries the source material section by section and assembles structured, readable documentation that preserves the important details instead of collapsing everything into a thin summary. It is aimed at a common pain: recorded calls and long documents that nobody has time to convert into proper written docs. Because it calls commercial LLM APIs under the hood, output quality tracks the model you choose, and you stay in control by guiding the outline before generation begins.
Lumentis is free and open source; there is no paid plan or subscription from the project. The only cost is usage of whichever LLM API you connect, such as OpenAI or Google, and that usage is billed directly by those providers at their normal rates.
Teams that record lots of calls and need searchable written documentation, founders turning customer interviews into structured notes, and writers converting raw transcripts into drafts. It suits anyone who prefers reviewing an outline first, then letting the AI write the full document.