This software provides a private environment for artificial intelligence tasks. You run everything on your own hardware. Your data never leaves your machine. This stack helps you organize documents, build knowledge bases, write code, and fine-tune models.
Local-ai-stack connects several tools into one package. It uses Ollama to handle the language models, FastGPT for your chat interface, and pgvector to search through your private documents. You do not need an account or an internet connection to use these tools.
Your computer needs sufficient resources to run these tools. We recommend:
Follow these steps to install the software on your Windows computer:
Once the application loads, you see a browser window. This is the control panel. From here, you can upload your PDF or text files. The system processes these files so you can ask questions about them. You can also select different AI models to help with code or general tasks.
If you want to create a knowledge base, navigate to the Document tab. Click the upload button to add your documents. The software will turn your files into a searchable index. You can then use the chat window to talk to your data.
This software operates offline. The connection between your computer and the AI tools stays local. No third party accesses your knowledge base or your files. Your work remains on your hard drive at all times.
Large models require more memory. If you experience slow responses, try selecting a smaller model within the settings menu. You can also limit the number of documents the system searches at one time to improve speed.
Keywords: ai-agent, docker, document-retrieval, fastgpt, fine-tuning, knowledge-base, llm, local-ai, lora, neural-search, nlp, ollama, open-source, private-llm, qwen, rag, retrieval-augmented-generation, self-hosted, text-embedding, vector-database