Internal manuals and processes
Query instructions and procedures without searching through folders. Retrieve the relevant passage and link to the document version appropriate for the user.
Use cases / Data and knowledge
Chat with documents
Build an AI chat that answers questions about your documents.
Ask questions about PDFs, contracts and manuals in natural language. Turn company documentation into an assistant that helps people find information and check its sources.
With QDivZero, you can combine semantic search, RAG and language models to build your own document chat. You connect the content and design the experience; QDivZero runs the compatible models.
Make your documents easier to use
Find the information without reading every page.
Turn your files into useful conversations. Help teams and customers find answers and consult their sources.
Contracts and PDF files
Ask about clauses, dates and terms in your PDFs. Show source passages and distinguish contracts, annexes and versions.
Product documentation
Resolve configuration questions using product manuals. Scope content by version and accompany answers with useful steps and links.
Document libraries
Query several collections from one interface. Filter by project, date or scope and retain references to each file.
Build it with QDivZero
Prepare your documents. Answer each question.
Index documents with references and retrieve passages for each question. An LLM prepares the answer; your application shows sources and limits the content each user can query.
Prepare your data
Documents
Split content into passages and keep source references.
Embeddings
Represent each passage using your chosen embedding model.
Vector index
Store vectors with content and metadata.
For each query
Question
The user asks about the content in natural language.
Retrieve context
Retrieval finds passages related to the question.
Generate the answer
The LLM receives the question and retrieved passages.
Show sources
Your application presents the answer and available references.
Open-weight / Hugging Face
You can start with…
A PDF chat combines an embedding model to retrieve context and an LLM to write the answer. Add a reranker when you need better passage selection, and compare each component with real questions about your documents. Test long documents, similar files, and queries requiring specific page references. A suitable combination should retrieve the right passage and produce an answer faithful to the available context.
Embeddings
Qwen3-Embedding-8B
Represent documents and queries as vectors for semantic similarity search.
View model on Hugging FaceReranking
Qwen3-Reranker-8B
Reorder retrieved results according to their relevance to the query.
View model on Hugging FaceText and vision
Qwen3.8-27B
For conversation, code, and tasks combining text, images, and your own context.
View model on Hugging FaceChat with documents
Chat with documents: frequently asked questions
What is AI document chat and how does it work?
AI document chat answers questions about PDFs, manuals, or company files. It retrieves passages related to each question and sends them to a language model as context. Combine QDivZero Retrieval and an LLM to build this experience.
How do I build an AI chatbot for PDF documents?
Prepare the PDF text, split it into passages, and generate embeddings for indexing. Connect a chat interface to Retrieval and the model answering questions. Your application coordinates queries and displays the result.
Can I ask questions across multiple documents?
Yes. Index several documents in a collection and retrieve information from different files for one query. Keep source metadata to identify which PDF, manual, or contract provides each passage used in the answer.
How do I show sources in a document chatbot?
Store the filename, page, and passage reference during indexing. Return that metadata alongside the answer when retrieving context. This lets you build a PDF chatbot that provides access to the original content.
Can I build a private chatbot for company documents?
Deploy compatible models and connect internal documents through Retrieval. QDivZero processes inference content without retaining prompts and responses or using them to train models. Your application manages document storage and determines what each user can query.
Which models do I need for document chat?
Use an embedding model to retrieve content and an LLM to generate answers. Start with Qwen3-Embedding and Qwen3.8, and add a reranker to refine context selection. Choose versions compatible with your deployment.
How do I keep a document chatbot up to date?
Update passages and embeddings when a file changes, and remove content that should no longer be searchable. Keep versions and source references so each question retrieves the right documentation. This lets you expand your PDF chat without retraining the language model whenever information changes.
How much does an AI chat with PDFs cost?
The cost combines document preparation, embedding generation, the index and LLM inference. File volume, concurrent questions and context length help determine the capacity you need. Check QDivZero pricing and start with a representative collection to choose a deployment that fits your application.
Ready to make your documents answer questions?
Build a chat for PDFs and internal documentation using your data and the models you choose. Combine Retrieval and an LLM in QDivZero, connect your application through an API and show answers people can check against the original sources.