● LIVEReading: Chat With Your Own Docs Using AnythingLLM DesktopTotal time: 8 minSteps: 5Worked first time: 82%● LIVEReading: Chat With Your Own Docs Using AnythingLLM DesktopTotal time: 8 minSteps: 5Worked first time: 82%
CBW
Easygithub.com/Mintplex-Labs/anything-llm2026-05-19

Chat With Your Own Docs Using AnythingLLM Desktop

Install AnythingLLM on your computer, connect a free AI model, and start chatting with your own PDFs and documents in under 15 minutes. No coding required.

// Build stats

  • Total time8 min
  • Number of steps5
  • DifficultyEasy
  • Worked first time82%
// Before you start

What you need

  • Windows, Mac, or Linux computer
  • At least 8 GB RAM (16 GB recommended for local AI models)
  • An OpenAI API key OR Ollama installed locally (for a free option)
  • A PDF, DOCX, or TXT file you want to chat with
  • Stable internet connection for the initial download
01
Step 1 of 5

Download and install the AnythingLLM desktop app

3 min

AnythingLLM offers a one-click desktop app for Mac, Windows, and Linux. This is the easiest way to get started — no Docker, no terminal, no configuration files. Just download and run the installer like any normal app.

Terminal · mac
$ Open your browser and go to: https://anythingllm.com/download
$
$ Download the installer for your operating system, then double-click it to install.
What you should see
The AnythingLLM app opens to a welcome/onboarding screen asking you to set up your LLM provider.
This might happen

On Mac you may see a warning that the app is from an unidentified developer.

Go to System Settings > Privacy & Security, scroll down, and click 'Open Anyway' next to the AnythingLLM entry.

02
Step 2 of 5

Choose and connect your AI model

3 min

AnythingLLM needs an AI model to power the chat. You have two main options: (A) Use OpenAI — fast and easy, costs a small amount per query, requires an API key from platform.openai.com. (B) Use Ollama — completely free and private, runs on your own machine, but requires Ollama to already be installed (ollama.com). During onboarding, pick whichever fits you. For OpenAI, paste your API key when prompted. For Ollama, make sure Ollama is running first, then select it and pick a model like llama3.

Terminal · mac
$ For OpenAI: Get your API key at https://platform.openai.com/api-keys
$
$ For Ollama (free/local): Install from https://ollama.com, then run in your terminal:
$ ollama pull llama3
What you should see
The onboarding screen shows a green checkmark or 'Connection successful' message after you save your LLM settings.
This might happen

Ollama connection fails even though Ollama is installed.

Make sure Ollama is actually running. Open a terminal and type 'ollama serve' to start it, then return to AnythingLLM and try again.

03
Step 3 of 5

Create your first workspace

1 min

A workspace is like a private chat room that holds a specific set of documents. Think of it as a folder for a project. You can have multiple workspaces — one for work contracts, one for research papers, etc. Each workspace only knows about the documents you put into it.

Terminal · mac
$ In the AnythingLLM sidebar, click the '+ New Workspace' button.
$ Type a name (e.g. 'My Documents') and click Create.
What you should see
A new workspace appears in the left sidebar and the chat window opens for it.
04
Step 4 of 5

Upload your documents to the workspace

2 min

Now you load in the files you want to chat with. AnythingLLM supports PDF, DOCX, TXT, and more. It reads the content, breaks it into chunks, and stores it so the AI can search through it when you ask questions. Drag and drop your file directly into the chat window, or use the upload button in the workspace settings.

Terminal · mac
$ Click the upload icon (paperclip or document icon) inside your workspace.
$ Drag and drop your PDF or DOCX file onto the upload area, or click to browse for it.
$ Click 'Move to Workspace' then 'Save and Embed'.
What you should see
A progress bar appears while the document is processed. When done, the file appears in the workspace document list with a green status indicator.
This might happen

Large PDFs (100+ pages) take a long time to embed or appear to hang.

This is normal — embedding is doing real work. Wait up to 5 minutes. If it truly freezes, restart the app and try uploading a smaller file first to confirm everything works.

05
Step 5 of 5

Start chatting with your documents

2 min

Once your document is embedded, just type a question in the chat box as if you were texting someone. Ask it to summarize, find specific facts, explain a clause, or compare sections. The AI will answer using your document as its source and show you which parts of the document it pulled from.

Terminal · mac
$ Type your question in the chat box at the bottom of the workspace, for example:
$ 'Summarize the key points of this document'
$ or
$ 'What does this document say about payment terms?'
$
$ Press Enter to send.
What you should see
The AI replies with an answer and shows source citations at the bottom of the response, referencing the specific sections of your document it used.
This might happen

The AI gives a generic answer and says it has no information about your document.

The document may not have been embedded properly. Go to workspace settings, check that your file shows a green status, and click 'Save and Embed' again if needed.

// Status

cooked. baked. worked.

A running desktop app where you can upload any PDF or document and have a real conversation with its contents, with source citations, using either a free local AI model or OpenAI.

// the honest bit

The honest part

The desktop app works well for personal, single-user use. Multi-user support, the embeddable chat widget, and some advanced agent features are only available in the Docker (self-hosted server) version. Local AI models via Ollama require a reasonably powerful machine — on older laptops with 8 GB RAM, responses can be slow. OpenAI gives faster results but costs money per query. AnythingLLM collects anonymous usage telemetry by default; you can disable this in Settings > Privacy.