AnythingLLM: Chat with Your PDFs Offline with Ollama (Free, No Account) @KGPTalkie
AnythingLLM: Chat with Your PDFs Offline with Ollama (Free, No Account)  @KGPTalkie
Uploaded October 2026 | Updated November 2026, 5 days ago
Set up AnythingLLM on Windows and chat with your own PDFs using a local Ollama model: every answer comes from your files, with the source shown. Switch the workspace to Chat mode and nothing leaves your computer. The default Agent mode searches the web with its Web Search skill and ignores your PDF, and this video shows the one-click fix. Day 9 of the Free AI Setup series: one free AI tool a day, set up step by step.

Watch the whole series in order: youtube.com/playlist?list=PLfUBzZXRNC-k
Previous, day 8, Codex CLI: youtu.be/J1Z_yf9lhgk
Next, day 10, OpenRouter free models: youtu.be/l5oBv_Mhf3k

What you need: a Windows computer (the docs recommend 16 GB of RAM and an 8-core processor), Ollama with one model (day 1 of this series), and nothing else. No account, no email, no credit card. A graphics card is optional: it speeds up big models, and a small model runs on the processor alone, only slower.

In this video:
- What AnythingLLM is: a free, open source desktop app, the filing cabinet and front desk for your documents, with Ollama as the brain
- Download (about 391 MB) and install for your user, no administrator prompt (the Meeting Assistant models, about 800 MB, are skipped)
- Setup screens: skip the built-in model, pick Ollama, and let the app find it at its default address, then choose your model
- Privacy and Data: switching Anonymous Telemetry off, and what the embedder (all-MiniLM-L6-v2, 23 MB as the app labels it), LanceDB and the chunk settings are
- Workspaces: upload two PDFs, Move to Workspace, Save and Embed, and why the first embed waits about 4 minutes for a one-time download
- The mistake to avoid: in the default Agent mode, the Web Search skill answers from the web and ignores your PDF
- The fix: Chat Settings, Chat mode, Chat; the real answer to an air fryer error code in about 2 seconds, with the exact page as the source
- The offline check: in Chat mode the app opened no connection except to this computer
- More questions: a router factory reset, and one answer that needs two documents
- Chat vs Query vs Agent: Query refuses a question your files do not cover
- Speed on an RTX 5090: Qwen 3.8 27B, about 16 GB of graphics memory, about 150 tokens per second
- Tips: a new thread per topic, the greeting name, adding more files later, free vs a paid chat service (about 20 USD a month)

Chapters:
0:00 Intro: ask your own PDF, answer in 2 seconds
0:31 Day 9 of 30 and what AnythingLLM is
1:09 What you need
1:36 Step 1: Download AnythingLLM
1:50 Step 2: Install
2:28 Step 3: Setup screens and Ollama
3:35 Privacy and Data, the embedder
4:21 Step 4: Workspace and your PDFs
5:28 Step 5: First question, the Agent mode gotcha
6:12 The fix: Chat mode
7:03 Offline check: only the local address
7:15 More questions: router reset and warranty
8:04 Chat vs Query vs Agent
8:56 Speed on an RTX 5090
9:29 Three quick tips
9:50 AnythingLLM vs a paid chat service
10:23 Next: OpenRouter free models

Links
AnythingLLM download: anythingllm.com/download
AnythingLLM docs, Windows desktop install: docs.anythingllm.com/installation-desktop/windows
AnythingLLM docs, system requirements: docs.anythingllm.com/installation-desktop/system-requirements
AnythingLLM docs, chat modes: docs.anythingllm.com/features/chat-modes
Ollama setup (day 1 of this series): ollama.com/download

Tomorrow, day 10: OpenRouter free models.

New free AI tool every day. Subscribe so you don't miss one: youtube.com/@KGPTalkie?sub_confirmation=1
AnythingLLM: Chat with Your PDFs Offline with Ollama (Free, No Account)10 Concatenation and Sorting - Numpy Crash Course for Data Science | Numpy for Machine LearningI Ran Google MedGemma 1.5 on My PC 🤯 Best Free Medical AI Yet?UPI Payment System DesignOrder State Machine: Delivered, Then Picked Up Again?Fable 5.2 vs Gemini 4 Argon: Same Score, 5x Cheaper?Quick and Simple Sentiment Analysis App in PythonNemotron 3.5 Lightning 33B vs Claude Haiku 5.5: Is Fast Also Good?NumPy Crash Course for Machine Learning | Arrays, Indexing, Operations and More02 RAGWire RAG Tutorial: Multi Company RAG for Financial Comparison using OllamaBPE vs WordPiece vs SentencePiece: Whats the Difference? | LLM InterviewAWS Tutorial 1 - Launching EC2 Ubuntu Machine on AWS
Laxmi Kant | KGP Talkie |

AnythingLLM: Chat with Your PDFs Offline with Ollama (Free, No Account)

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER