{% extends "base.html" %} {% block title %}ArtificeGraph — About{% endblock %} {% block content %}
Local-first · offline · print-inflected

About ArtificeGraph

ArtificeGraph is a modular, local-first pipeline that transforms historical OCR text into a structured knowledge graph and a hyperlinked Obsidian vault — entirely offline using local LLMs and embedding models. The web interface mirrors the tkinter GUI: configure paths and the LLM endpoint, then run each stage or the whole pipeline at once.

Pipeline stages

  1. Ingest — discovers .txt/.md files and produces sliding-window chunks.
  2. Extract — sends batches to a local Ollama or OpenAI-compatible LLM, parsing structured JSON into entities and relationships.
  3. Resolve — fuzzy & (optionally) bge-m3 semantic dedup with YAML alias overrides.
  4. Vault — emits an Obsidian vault with Dataview frontmatter and wikilinks.
  5. Graph — exports GraphML / GEXF / JSON / CSV for Gephi, yEd, and beyond.

Design language

This interface implements the The New Masses design system: an offline-first, print-inflected editorial aesthetic with warm paper surfaces, Esperanto-green accent, and fluid serif typography. All design tokens are served from the shared packages/shared-ui/shared_ui/assets/tokens.css; never hardcode colours or spacing.

Run it

From the project root: 

pip install -e ".[web]"
python -m web.server
# → http://localhost:8766

Vanilla JS · IIFE-wrapped · no framework · no transpiler.

{% endblock %}