ForkFree

Stacks

Build an AI assistant that answers from your documents

Turn your manuals, policies and FAQs into a searchable assistant for staff or customers.

For a small team that answers the same questions repeatedly. RAGFlow reads and indexes your documents, Ollama runs the model locally, and Chatwoot or AnythingLLM puts the answers in front of people.

  1. 1

    ragflow

    Read PDFs, Word files and spreadsheets and index them for question answering

    Instead of . Apache-2.0 Worth a look

  2. 2

    ollama

    Run the language model on your own hardware

    Instead of . MIT Worth a look

  3. 3

    anything-llm

    Give staff a simple chat screen over the same documents

    Instead of . MIT Worth a look

  4. 4

    chatwoot

    Handle customer chats and hand over to a person when needed

    Instead of Intercom. MIT Worth a look

  5. 5

    Stirling-PDF

    Fix, split and OCR messy PDFs before indexing

    Instead of Adobe Acrobat Pro. MIT Ready to use

Before you start

['Local models give weaker answers than top paid ones, and may state wrong things confidently. Have staff check anything important.', 'RAGFlow is heavy to run and wants a server with plenty of memory.', 'Do not put the assistant in front of customers until you have tested it on real questions.']