Find your next open-source project

Search by name, purpose, or keyword — or paste a GitHub repository URL

shnfatih-hash/devlet-arsivleri-soykutugu

Learning Resources

An agent skill package for tracing Ottoman and Turkish ancestry through official state archives, from 1500s land registers to e-Devlet records.

This repo is an agent skill pack · 1 skillsSkill list and install →
108
9
Python
MIT
new today

What It Is

This repository packages a research skill for tracing Turkish and Ottoman family lines below the point where the e-Devlet Alt-Üst Soy Bilgisi (ancestry) document stops. It combines a SKILL.md instruction set for AI agents with a step-by-step archive methodology covering Ottoman census registers, tax records, court (kadı) registers and 16th-century land surveys. The README explains why ancestry documents cut off at different dates, charts which register series (NFS.d, MAD.d, TT.d, MŞH.ŞSC.d) cover each period back to the 1500s, and lists the official portals where each is searched catalogued registers and portals. It is aimed at hobbyist and semi-serious genealogists, not at developers building software.

The intended setup is an agent environment such as Google Antigravity, Claude Code or Codex, connected to the user's logged-in Chrome session via the official Chrome DevTools MCP bridge, so the agent can search the archive catalogue without receiving the user's credentials. A Python/Pillow script crops downloaded manuscript page images into 1:1 pixel sections for reading, and a helper JavaScript file sequences catalogue queries.

Getting Started

  1. Download your e-Devlet ancestry document and, for immigrant families, the name-equivalence certificate. Note the oldest ancestors, birth years and the village's old name.
  2. Install Python 3 with Pillow, Node.js and Chrome, then clone the skill into your agent's skills directory. For Claude Code or Codex the command is git clone <repo> ~/.claude/skills/devlet-arsivleri-soykutugu; Antigravity uses ~/.gemini/config/skills/ installation commands.
  3. Add the Chrome DevTools MCP server configuration to your agent's MCP config (npx -y chrome-devtools-mcp@latest --autoConnect), enable remote debugging at chrome://inspect/#remote-debugging, and log in to the state archives catalogue with e-Devlet yourself.
  4. Have the agent check the register's index (Mi=1..5) first and report the correct page range, then buy and download only those pages from e-Satış.
  5. Open the downloaded .zip/.jpg folder in the agent and send the main prompt, which asks the agent to verify each key entry with pixel crops and label findings [KESİN], [GÜÇLÜ] or [İSİM İZİ] quick-start flow.

Constraints

Searching the archive catalogue is free, but high-resolution watermark-free page images are sold per image through e-Satış; the README explicitly warns against buying an entire 300–500-page register blindly and recommends an index test first cost and index test. The skill assumes a logged-in browser session; credentials are never handed to the model. Parallel requests to the catalogue can trigger a HTTP 418 block, so searches should be sequential. The crop_and_verify.py helper requires Python 3 and Pillow. The package is MIT-licensed package layout.

HotRepo's Take

On the day it entered HotRepo's Python new-project list, the repository recorded 102 new stars and 9 forks in roughly ten and a half hours, with no all-language daily rank recorded. Its 102 stars are almost entirely from that first window, so the data supports only a launch-week observation, not a trend. What stands out is the format: a genealogy workflow expressed as an agent skill, with domain documentation (register series, Ottoman terminology, immigrant-family routing) doing most of the work rather than code, which fits a niche that sits between family-history research and tooling guidance.

Written by AI from the project's public information and HotRepo's records, updated . The project's own docs are the source of truth.