AI · Desktop app

AnythingLLM

Private, on-device AI assistant to chat with documents and local models.

Visit anythingllm.com ↗

AnythingLLM is an open-source, local-first AI application that allows you to run language models directly on your hardware without external API keys or token limits. It bundles document ingestion, vector storage, web search, and agent skills into an all-in-one desktop or self-hosted workspace. It is especially useful for querying private files, meeting notes, and internal data without sending text to the cloud.

What it's for

  • Chat privately with local documents like PDFs, spreadsheets, and Word files without cloud uploads
  • Record, transcribe, and summarize meetings locally without inviting third-party bots
  • Run open-source LLMs locally on CPU, GPU, or NPU hardware
  • Self-host a shared AI workspace for team document querying using Docker

Features

  • All-in-one local installation with built-in embeddings and vector storage
  • Document ingestion across common formats including PDF, Word, and spreadsheets
  • Local meeting assistant that records, diarizes, and generates summaries on-device
  • Custom agent skills and background job scheduling
  • Integrated web scraping and real-time web search
  • Multi-user deployment support via Docker and developer APIs

Pricing

Open source. Free and open source (MIT license); optional hosted Cloud version available

Details

Type
Desktop app
Category
AI
Topic
Local LLM Apps
Open source
Yes
Platforms
macOS, Windows, Linux, Android, Self-hosted
For
Privacy-conscious professionals, developers, and organizations running local AI
Alternative to
LM Studio, Ollama, GPT4All, Granola
Tags
llm, rag, local ai, privacy, desktop app, open source

Why it was saved

  • @techfren, 17 July 2024: It offers an all-in-one private alternative to cloud chatbots and GPT4All, featuring one-click install, multi-format file ingestion (PDF, Word, spreadsheets), built-in embeddings, data scrapers (GitHub, YouTube, Confluence), and agent capabilities.

Related

Updated