gitwtfhub

wtf is doc7?

magicrew/doc7 — explained in plain English

Analysis updated 2026-08-13

1,095GoAudience · developerComplexity · 3/5LicenseSetup · moderate

TL;DR

A command line tool that uses a vision AI model to convert PDFs, scans, and Office files into AI-searchable Markdown, even on pages with no text layer.

Mindmap

mindmap
  root((doc7))
    What it does
      Documents to Markdown
      Visual page understanding
      No required OCR
    Tech stack
      Go
    Use cases
      Convert PDFs and scans
      Extract charts and tables
      Local AI chat
    Audience
      Developers
      Local AI users

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Why would anyone build with this?

REASON 1

Convert PDFs, scans, and Office files into AI-searchable Markdown.

REASON 2

Extract tables, charts, formulas, and diagrams from pages with no text layer.

REASON 3

Chat with a local AI model and ask it to convert specific documents.

REASON 4

Run document conversion entirely locally without a cloud OCR service.

What's in the stack?

Go

How it stacks up

magicrew/doc7nianzhibai/91versenilvis/iris
Stars1,0951,1991,217
LanguageGoGoGo
Last pushed2026-07-03
MaintenanceMaintained
Setup difficultymoderatemoderateeasy
Complexity3/53/52/5
Audiencedevelopergeneraldeveloper

Figures from each repo's GitHub metadata at analysis time.

How do you spin it up?

Difficulty · moderate Time to first run · 30min

Requires a local or remote OpenAI-compatible vision model, such as one running in LM Studio or Ollama.

Use, modify, and distribute freely, including for commercial purposes, as long as you keep the copyright notice.

Wtf does this do

doc7 is a command line tool that converts documents into clean, well organized text that AI tools can search and reason over. It works with PDFs, Office files like Word documents, scanned pages, screenshots, charts, formulas, and diagrams, turning all of them into a single text format called Markdown, which is a simple way of writing formatted text using plain characters. Unlike traditional document conversion tools that mainly extract text character by character, doc7 uses a vision capable AI model to actually look at each page the way a person would, reading tables, understanding chart relationships, and interpreting diagrams, even on pages that have no selectable text at all, such as a scanned image. You bring your own AI model, either running locally through tools like LM Studio or Ollama, or a remote service, as long as it speaks the common OpenAI compatible format. There is no required text recognition software and no per page fee charged by doc7 itself. The project includes a benchmark comparing its output against two other document conversion tools on a set of scanned pages with fifteen checkable facts. In that test, doc7 correctly recovered all fifteen facts, while the two comparison tools recovered nine and three respectively. Beyond simple conversion, doc7 also includes a chat mode, a small assistant running on your configured local model that can hold a conversation and, when you explicitly ask it to process a file, convert that file into Markdown as part of the chat. It can also help set up your local model configuration in plain language and includes safety limits so it can only browse a restricted set of folders and never runs arbitrary system commands. Installing doc7 does not require administrator privileges and works on macOS, Linux, and Windows. Once installed, it automatically finds any local vision model server already running on your machine and lets you pick which model to use. This project is licensed under the MIT license, a permissive license that allows free use, modification, and commercial use. doc7 is aimed at developers and technical users who want to feed documents into AI systems, particularly those who prefer running everything locally rather than sending files to a cloud service.

Yoink these prompts

Prompt 1
Help me install doc7 and connect it to a local vision model in LM Studio.
Prompt 2
Explain how doc7's visual understanding differs from traditional OCR.
Prompt 3
Show me how to use doc7 chat mode to convert a report into Markdown.
Prompt 4
Walk me through configuring doc7 to use a remote OpenAI-compatible endpoint.

Frequently asked questions

wtf is doc7?

A command line tool that uses a vision AI model to convert PDFs, scans, and Office files into AI-searchable Markdown, even on pages with no text layer.

What language is doc7 written in?

Mainly Go. The stack also includes Go.

What license does doc7 use?

Use, modify, and distribute freely, including for commercial purposes, as long as you keep the copyright notice.

How hard is doc7 to set up?

Setup difficulty is rated moderate, with roughly 30min to a first successful run.

Who is doc7 for?

Mainly developer.

View the repo → Decode another repo

This repo across BitVibe Labs

Don't trust strangers blindly. Verify against the repo.