Self-hosted document management system that scans, OCR-indexes and archives paper documents into a searchable online archive.
icereed/paperless-gpt
Uses LLMs and vision OCR to generate titles, tags, correspondents and custom fields for documents in paperless-ngx.
- Stars
- 2,821
- Forks
- 235
- Open issues
- 190
- Last commit
- yesterday
What is it?
paperless-gpt is a companion service for paperless-ngx, the self-hosted document management system. It sends document content to an LLM such as OpenAI, Mistral, Anthropic or a local Ollama model, then writes back suggested titles, tags, correspondents, document dates and custom fields. Users can review suggestions in a web UI or let documents be processed automatically using a trigger tag.
Its OCR path uses vision-capable LLMs to extract text from scans, and it can also use Azure Document Intelligence, Google Document AI or a self-hosted Docling server. Prompts are editable from the web interface, and workflows let each document type use its own trigger tag and prompts. The project also links documents that cite the same invoice, contract or case number through paperless-ngx Document Link fields.
The service is deployed with Docker, typically alongside paperless-ngx in the same Compose file. It has no built-in authentication, so the README recommends placing it behind a reverse proxy or a VPN.
Key features
- 01Generates titles, tags, correspondents and created dates from document content using LLMs.
- 02Extracts text from scans with vision LLMs, Azure Document Intelligence, Google Document AI or Docling.
- 03Supports OpenAI, Mistral, Anthropic and local Ollama models.
- 04Fills custom fields in append, update or replace modes.
- 05Lets users edit prompts and define per-document-type workflows in the web UI.
Use cases
- Organizing a large backlog of scanned paperwork in paperless-ngx without manual tagging.
- Digitizing poor-quality scans where traditional OCR gives weak results.
- Linking invoices, reminders and contracts that reference the same document number.
- Running document processing with a fully local Ollama model to keep data on your own hardware.
Quick start
- 1
Install Docker
- 2
Prepare a running paperless-ngx instance and an LLM provider (OpenAI API key or Ollama server)
- 3
Create a docker-compose.yml based on the README example, alongside paperless-ngx
- 4
Start the services with Docker Compose
Things to know
- paperless-gpt has no built-in authentication, so it must be placed behind a reverse proxy or restricted network.
- It needs an LLM provider, such as an OpenAI API key or a running Ollama server.
In collections
Topics
FAQ
- What is paperless-gpt?
- Uses LLMs and vision OCR to generate titles, tags, correspondents and custom fields for documents in paperless-ngx. paperless-gpt is a companion service for paperless-ngx, the self-hosted document management system. It sends document content to an LLM such as OpenAI, Mistral, Anthropic or a local Ollama model, then writes back suggested titles, tags, correspondents, document dates and custom fields. Users can review suggestions in a web UI or let documents be processed automatically using a trigger tag.
- Is paperless-gpt free to use? What license does it have?
- paperless-gpt is open source under the MIT license. Check the license terms before using it in commercial products.
- Is paperless-gpt still maintained?
- Actively maintained. The last commit to paperless-gpt was on Nov 6, 2026.
- What can I use paperless-gpt for?
- Organizing a large backlog of scanned paperwork in paperless-ngx without manual tagging. Digitizing poor-quality scans where traditional OCR gives weak results. Linking invoices, reminders and contracts that reference the same document number. Running document processing with a fully local Ollama model to keep data on your own hardware.
- How do I install paperless-gpt?
- Install Docker Prepare a running paperless-ngx instance and an LLM provider (OpenAI API key or Ollama server) Create a docker-compose.yml based on the README example, alongside paperless-ngx Start the services with Docker Compose
- Where can I get paperless-gpt?
- The source code and installation instructions are on GitHub: https://github.com/icereed/paperless-gpt
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Papra is a minimalistic, self-hostable document management and archiving platform for long-term storage of personal and shared documents.
Self-hosted document management system for scanned documents, with OCR, full-text search, tags, folders and a desktop-style web UI.
Self-hosted personal document management system that ingests scans, emails and files, runs OCR and suggests metadata with machine learning.
Self-hosted accounting app that uses LLMs to extract data from receipts, invoices and PDFs into a structured transaction database.
A JetBrains IDE plugin that generates git commit messages from your diff using a configurable LLM backend.