Private AutoGPT Robot - Your private task assistant with GPT!
- 🔥 Chat to your offline LLMs on CPU Only. 100% private, no data leaves your execution environment at any point.
- 🔥 Ask questions to your documents without an internet connection. Engine developed based on PrivateGPT.
- 🔥 Easy coding structure with Next.js and Python. Easy to understand and modify.
- 🔥 Built with LangChain, GPT4All, Chroma, SentenceTransformers, PrivateGPT.
The supported extensions are:
.csv: CSV,.docx: Word Document,.doc: Word Document,.enex: EverNote,.eml: Email,.epub: EPub,.html: HTML File,.md: Markdown,.msg: Outlook Message,.odt: Open Document Text,.pdf: Portable Document Format (PDF),.pptx: PowerPoint Document,.ppt: PowerPoint Document,.txt: Text file (UTF-8),
- Python 3.8 or higher.
- Install PAutoBot:
pip install pautobot- Run the app:
python -m pautobot.appor just:
pautobot- Go to http://localhost:5678/ to see the user interface. You can choose one of the two modes:
- Chat Only
- Documents Q&A
- Upload some documents to the app (see the supported extensions above). You can try docs/python3.11.3_lite.zip for a quick start. This zip file contains 45 files from the Python 3.11.3 documentation.
- Force ingesting documents with Ingest Data button.
You can also run PAutoBot publicly to your network or change the port with parameters. Example:
pautobot --host 0.0.0.0 --port 8080git clone https://github.com/nrl-ai/pautobot
cd pautobot- Python 3.8 or higher.
- To install Pautobot from source, from
pautobotsource code directory, run:
pip install -e .- Run the app:
python -m pautobot.app- Go to http://localhost:5678/ to see the user interface.
- Install the dependencies:
cd frontend
npm install- Run the app:
npm run dev- Go to http://localhost:3000/ to see the user interface. Use this address to develop the frontend.