PDF-ChatGPT

About

PDF-ChatGPT is a free, open-source application for asking questions about PDF documents and receiving answers with page citations. It runs self-hosted, indexes multiple PDFs in a persistent FAISS index, and lets users scope semantic search to a document. Answers use the OpenAI Responses API and can stream token by token; SQLite stores conversation history for follow-up questions. The browser interface renders Markdown, offers dark and light themes, and adapts to mobile and tablet layouts. The project is built with FastAPI, LangChain, FAISS, and OpenAI’s Responses API. PDFs, the index, and conversation history are stored locally, while text is sent to OpenAI for embeddings and chat generation. An OpenAI API key with billing enabled is required, and token usage is metered. Uploads accept PDFs under a configurable limit, with a listed maximum of 25 MB. Scanned PDFs without extractable text need OCR, which is not included. The default setup has no login and is single-tenant. Local operation and deployment to several hosting platforms are documented; persistent storage is needed to retain data across redeploys.

Who it is for

It suits developers or teams who want a self-hosted PDF question-answering app with source-page citations and can use OpenAI APIs. It is a poor fit for scanned PDFs unless they add OCR, or for public deployment without adding access controls.

Compared on AI PDF assistants

Maximum file size

25 MBgithub.com

Multiple PDFs per chat

Yesgithub.com

Source citations

Yesgithub.com

Scanned PDF support

Nogithub.com

Export formats

[]github.com

Access platforms

Web, Windows, macOS, Linux, APIgithub.com

Facts

Purpose

PDF-ChatGPT is a self-hosted RAG application for asking natural-language questions about PDF documents and receiving answers with page citations.github.com · 7 Oct 2026

Document library

It supports uploading and indexing multiple PDFs and shows document status, page counts, and chunk counts.github.com · 7 Oct 2026

Answer generation

The app uses OpenAI's Responses API and can stream answers token by token.github.com · 7 Oct 2026

Search

It uses semantic search over a persistent FAISS index, with a relevance threshold and optional per-document scoping.github.com · 7 Oct 2026

Citations

Answers include citations to the source document and page.github.com · 7 Oct 2026

Conversation features

Conversation history is stored in SQLite and prior turns are used for follow-up questions.github.com · 7 Oct 2026

Interface

The browser interface supports Markdown rendering, dark and light themes, and responsive layouts for mobile and tablet.github.com · 7 Oct 2026

Integrations

Out of the box, the project uses OpenAI embeddings and the OpenAI Responses API; the maker says adapting it to another LLM provider would require code changes.github.com · 7 Oct 2026

Usage costs

Users need an OpenAI API key with billing enabled, and embedding and chat usage is metered per token.github.com · 7 Oct 2026

Data handling

The maker says PDFs, the FAISS index, and conversation history are stored locally, while text is sent to OpenAI for embeddings and chat generation.github.com · 7 Oct 2026

Access control

The default setup is single-tenant with no login system, and the maker advises against exposing it publicly without authentication or network restrictions.github.com · 7 Oct 2026

File limits

Uploads are restricted to PDF files under a configurable size limit, and scanned PDFs without extractable text are not supported without adding OCR.github.com · 7 Oct 2026

Deployment

The maker documents local running and deployment on hosting platforms including Render, Railway, Fly.io, and Azure App Service; persistent storage is needed to retain data across redeploys.github.com · 7 Oct 2026

License

The repository is released under the MIT License.github.com · 7 Oct 2026

RAG stack

The application is built with FastAPI, LangChain, FAISS, and the OpenAI Responses API.github.com · 8 Oct 2026

PDF library

It supports uploading and indexing multiple PDFs and shows document status, page counts, and chunk counts in a sidebar library.github.com · 8 Oct 2026

Chat features

Conversation history is stored in SQLite, prior turns are used for follow-up questions, and answers can stream token by token over Server-Sent Events.github.com · 8 Oct 2026

Integration

The app requires an OpenAI API key and uses OpenAI embeddings; the README says API usage is metered per token and requires billing to be enabled.github.com · 8 Oct 2026

Storage

Uploaded PDFs and the FAISS index are stored locally, while conversation and message history is persisted in SQLite.github.com · 8 Oct 2026

Security setup

The README advises configuring the OpenAI key and other settings through environment variables or deployment secrets, and recommends TLS termination through Nginx or Caddy.github.com · 8 Oct 2026

Notable limits

Scanned image PDFs without an extractable text layer are unsupported without adding OCR, and the README lists OCR as a future improvement.github.com · 8 Oct 2026

API and setup help

FastAPI provides autogenerated Swagger, ReDoc, and OpenAPI documentation, and the repository includes a beginner's walkthrough.github.com · 8 Oct 2026

Where it ranks on Inferse

• Best AI PDF Assistants in 2026#8 of 27

Sources

• github.com/OlehDatsyk/pdf-chatgpt· checked 7 Oct 2026
• github.com/OlehDatsyk/pdf-chatgpt/blob/main/INSTRU· checked 7 Oct 2026