# AnythingLLM

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

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.

Website: https://anythingllm.com/

## What it is 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](https://xiry.app/topics/ai.md)
- 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](https://www.instagram.com/reel/C9g6hTMSF-H/): 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

- [GPT4All](https://xiry.app/resources/gpt4all): Desktop app to run open-source language models locally and completely offline (Desktop app, Open source)
- [Pinokio](https://xiry.app/resources/pinokio): Browser that installs and runs open-source AI apps with one click (Desktop app, Free)
- [Dolphin 3.0](https://xiry.app/resources/dolphin-3-0): Uncensored, open-source fine-tuned LLM designed for local execution without guardrails (AI model, Open source)
- [Stop Paying the OpenAI Tax: The Emerging Open-Source AI Stack](https://xiry.app/resources/stop-paying-the-openai-tax-the-emerging-open-source-ai-stack): Architectural guide breaking down developer-friendly open-source alternatives across modern AI layers (Article or course, Free)
- [Hermes 3](https://xiry.app/resources/hermes-3): Open-weights fine-tune of Llama 3.1 built for agentic function calling. (AI model, Open source)
- [8 Ways to Improve RAG Applications](https://xiry.app/resources/8-ways-to-improve-rag-applications): Practical architectural strategies to fix common failure modes in RAG systems (Tip, Free)

Updated 3 October 2026. Page: https://xiry.app/resources/anythingllm

Make your own library: Xiry takes your saved reels, tweets, Reddit bookmarks and Watch Later videos, makes them searchable and pulls the useful stuff out, like this library. It's invite-only for now; join the waitlist at https://xiry.app/#waitlist.