Page-1

# Set up AnythingLLM for local document Q&A AnythingLLM is an open-source AI application that provides a workspace for interacting with the model and adding your documents to provide context for the model responses. Instead of asking a model questions based only on its training data, AnythingLLM lets you ask questions about your own files and data. Think of it as a layer that sits on top of an LLM (such as Llama, Phi, GPT, DeepSeek, Mistral, Alibaba) and that provides the built-in features shown in the following table without requiring custom code. - Supports custom models: Run local models with the built-in LLM provider or connect to supported cloud providers, such as OpenAI, Azure, and AWS. - Use every type of document: Attach or embed documents of all types including PDFs, word documents, CSV, and codebases. You can even import docs from online locations. - Full privacy: AnythingLLM comes with sensible and locally running defaults for your LLM, embedding model, vector database, storage, and agents. Nothing is shared unless you allow it. - AI agents: Deploy autonomous AI agents that can reason, use tools, perform tasks, and execute multi-step workflows. - Vector database integrations: Connect with LanceDB, Chroma, Pinecone, Qdrant, Weaviate, and other vector stores for scalable retrieval. ## Get started Use the following steps to install AnythingLLM on Windows Arm and complete the initial setup. 1. [Download](https://cdn.anythingllm.com/latest/AnythingLLMDesktop-Arm64.exe) the latest version of AnythingLLM for Windows (Arm-compliant). 2. After it downloads, select and run the installer. 3. After the installation completes, open AnythingLLM. 4. Select **Manual set up** as the set up option. 5. Select your LLM preference. Depending on the LLM provider you select, the application prompts you for additional information. 6. Download the required model locally and select **Next**. After setup completes, AnythingLLM opens the workspace interface where you can create a workspace, add documents, and start chatting with the selected model. For more information about how to get started with AnythingLLM, see the [AnythingLLM documentation](https://docs.anythingllm.com/). ## Related - [API access and keys](https://docs.anythingllm.com/features/api) - [All features of AnythingLLM](https://docs.anythingllm.com/features/all-features) Last Published: Sep 15, 2026 [Previous Topic Local LLMs](https://docs.qualcomm.com/bundle/publicresource/80-62010-1/topics/llm-platforms.md) [Next Topic Microsoft AI Toolkit](https://docs.qualcomm.com/bundle/publicresource/80-62010-1/topics/ai_toolkit.md)