Local desktop and self-hosted AI workspace for document chat, retrieval, agents, and connections to local or hosted models.
Official sources reviewed Oct 4, 2026 · Some details not confirmed
What runs locally
Free desktop and self-hosted software; local models avoid provider token quotas while using your own compute.
Plan type
Free local software
Free allowance
Free desktop and self-hosted software; local models avoid provider token quotas while using your own compute.
Download on free
Not applicable
Watermark on free
Not applicable
Commercial use on free
Not confirmed
Credit card required
No
Paid plan
Optional hosted Basic: $50/month for a private instance; bring your own LLM API key and pay its provider separately.
READ THIS BEFORE YOU START
The limitation that matters
MIT-licensed application code does not grant universal rights to connected models or outputs. Desktop is local; hosted multi-user service and external APIs have separate costs.
A fully local setup needs local language and embedding models plus sufficient hardware. Official guidance recommends 16 GB RAM and an eight-core CPU for a basic local experience, with actual needs driven by the model. External models, databases, and services can add charges.
Plan details can change. A source check confirms published information; it is not a hands-on quality review or a guarantee that the plan fits every project.
Check the original sources
These are the provider pages used for this record. An unconfirmed field means we could not establish a clear answer from the reviewed information.
The software is free to run on your own computer. Free desktop and self-hosted software; local models avoid provider token quotas while using your own compute.
How much does AnythingLLM cost?
Optional hosted Basic: $50/month for a private instance; bring your own LLM API key and pay its provider separately.
Do I need a credit card to start using AnythingLLM?
No. The reviewed official sources say no credit card is required to start.
What is the main limit of AnythingLLM's free plan?
MIT-licensed application code does not grant universal rights to connected models or outputs. Desktop is local; hosted multi-user service and external APIs have separate costs.
Desktop app and Python library for running downloaded language models and chatting with local documents.
Free local software
Free local model inference on your hardware, with no hosted prompt-credit quota.
↳Private document chat on a personal computer
Keep in mindThe app's permissive software license does not license every model or its outputs. Check each model's commercial terms and hardware requirements.
Open-source desktop assistant for downloading local models, building assistants, connecting tools, and serving a local API.
Free local software
Free local inference after downloading a model; no cloud account or per-prompt usage fee is required.
↳Local AI chat with optional cloud providers and tool connections
Keep in mindJan's Apache 2.0 code license is separate from downloaded model licenses. Output-use rights and hardware requirements vary by model; connected online tools require network access.