A desktop app for saving, searching, graphing and chatting with your web pages, documents and files, with a built-in Chromium browser.
amosblomqvist/learnharness
A desktop app that uses AI agents to diagnose what you already understand and teach you toward a goal, with a persistent knowledge graph.
- Stars
- 5,668
- Forks
- 892
- Open issues
- 78
- Last commit
- 3 days ago
What is it?
LearnHarness is a desktop app for learning difficult subjects with AI agents. Before teaching, it measures where your understanding ends for each prerequisite of the lesson, then builds a dependency map from that starting point to your goal. Each step builds on what you already understand: you make predictions, receive explanations, and check your understanding, with confusions repaired before new material depends on them.
You can explain your reasoning, uncertainties, and reactions in your own words, by typing or dictation. Separate practice sessions pose varied problems that require the relevant knowledge to come to mind, with hints adapted to your reasoning. The app can also research questions in the background, generate visual explanations, and build interactive demonstrations.
What you learn is stored as propositions and reasoning chains in a knowledge graph that links back to source sessions, and you can import notes. Future sessions can build on this graph. Chats, graph data, settings, and API keys stay on your machine, while model requests go to the providers you configure, and the app supports 30+ provider options.
Key features
- 01Diagnoses your current understanding before each lesson and plans a dependency path to your goal.
- 02Lets you explain your reasoning by typing or dictating, and adapts teaching to your answers.
- 03Runs separate practice sessions with varied problems and adaptive hints.
- 04Stores learned ideas in a persistent knowledge graph linked to source sessions.
- 05Supports 30+ model provider options, including OpenAI, Anthropic, Google, DeepSeek, and OpenRouter.
- 06Keeps chats, graph, settings, and API keys on your local machine, with system prompts visible and editable.
Use cases
- Learning a dense technical subject where you need prerequisites filled in first.
- Studying a difficult topic over many sessions with a persistent record of progress.
- Importing your own notes so the agent can build on what you already wrote.
- Practicing recall of concepts you have been taught, without being told the terms first.
Quick start
- 1
Download the app from learnharness.ai or the GitHub Releases page
- 2
Add your provider API key and choose your models in the app
- 3
To run from source, install dependencies
npm install - 4
Start the development server at http://localhost:3000
npm run dev
Things to know
- Model costs are paid to your own AI providers, which requires an API key.
- Windows installer is unsigned and may trigger SmartScreen warnings.
- Licensed under AGPL-3.0, a copyleft license that requires sharing modified source when distributed or served over a network.
- Models, web research, and dictation send data to external services, so local storage does not mean offline processing.
In collections
Topics
FAQ
- What is learnharness?
- A desktop app that uses AI agents to diagnose what you already understand and teach you toward a goal, with a persistent knowledge graph. LearnHarness is a desktop app for learning difficult subjects with AI agents. Before teaching, it measures where your understanding ends for each prerequisite of the lesson, then builds a dependency map from that starting point to your goal. Each step builds on what you already understand: you make predictions, receive explanations, and check your understanding, with confusions repaired before new material depends on them.
- Is learnharness free to use? What license does it have?
- learnharness is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
- Is learnharness still maintained?
- Actively maintained. The last commit to learnharness was on Nov 4, 2026.
- What can I use learnharness for?
- Learning a dense technical subject where you need prerequisites filled in first. Studying a difficult topic over many sessions with a persistent record of progress. Importing your own notes so the agent can build on what you already wrote. Practicing recall of concepts you have been taught, without being told the terms first.
- How do I install learnharness?
- Download the app from learnharness.ai or the GitHub Releases page Add your provider API key and choose your models in the app To run from source, install dependencies: npm install Start the development server at http://localhost:3000: npm run dev
- Where can I get learnharness?
- The source code and installation instructions are on GitHub: https://github.com/amosblomqvist/learnharness
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A free, open-source desktop Cowork app that runs a built-in AI agent and coordinates multiple CLI agents like Claude Code and Codex.
A community-maintained native desktop app for installing, configuring, and chatting with the Hermes Agent AI assistant.
Cherry Studio is a cross-platform desktop AI client that connects to many LLM providers, local models and 300+ preset assistants.
Free, open-source desktop app where AI agents work on your local files, built on OpenCode and usable with any model provider or local Ollama models.
Open-source AI agent client for desktop and mobile that runs Claude Code, Codex and Pi across your own devices.