selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

GitHub
ArkhamMirror logo

Self-host ArkhamMirror

2.1.0

Local-first document intelligence for investigative analysis

ReleaseOpen SourceMIT
Updated9 months ago
LanguagePython
Setup difficultyAdvanced
Stars
498
Port
8100 (Main application API and UI)
Requires
pgvector, PostgreSQL
Install with Docker Compose
Dashboard with LLM configurationACH matrix with AI assistantForce-directed graph visualizationGeospatial graph visualizationTimeline analysis viewPattern detection viewCredibility assessment viewSearch with regex presetsMedia forensics analysis

What is ArkhamMirror?

SHATTERED is a modular platform for analyzing documents and conducting investigative research. It uses a plug-and-play architecture called Voltron with an immutable core Frame providing 17 services (database, vector search, LLM integration, job queues, event bus) and 26 self-contained shards for ingestion, extraction, organization, analysis, and export. Key capabilities include AI-powered analysis across all shards for anomaly detection, contradiction finding, credibility assessment, and summarization; full Analysis of Competing Hypotheses (ACH) with premortem, cone of plausibility, and devil's advocate mode; 10+ graph visualization modes including force-directed, hierarchical, geographic, and i2-style link analysis; timeline analysis with event extraction, conflict detection, and phase management; document processing pipeline supporting PDF, DOCX, images, HTML with OCR (PaddleOCR and vision LLMs), entity extraction, and claim extraction; hybrid semantic/keyword search with faceted filtering; and export to JSON, CSV, PDF, DOCX with Jinja2 templating. It runs via Docker Compose (single container with PostgreSQL) or manual Python/Node installation, requires PostgreSQL 14+ with pgvector, and supports fully air-gapped deployment with local LLMs (LM Studio, Ollama, vLLM). Authentication uses JWT with role-based access (Admin, Analyst, Viewer). Licensed MIT.

Should you self-host ArkhamMirror?

ArkhamMirror scores 50/100 on selfhost.directory (how scoring works) and ranks #28 of 46 in Document Management. It has 498 GitHub stars.

It shipped 3 stable releases in the past year. The latest, 2.1.0, came out on 18 January 2026. New versions land roughly every 3 weeks.

Setup is advanced: plan for several services and some ongoing maintenance. It is MIT-licensed and written mostly in Python.

Its closest alternative on selfhost.directory is Stirling-PDF (87/100, 95 252 stars). It scores higher overall, so compare the two before you pick. See ArkhamMirror vs Stirling-PDF side by side.

Features

  • AI-powered document analysis
  • Air-gap deployment
  • Analysis of Competing Hypotheses (ACH)
  • Anomaly Detection
  • Contradiction Detection
  • Credibility assessment
  • Document processing pipeline
  • Graph visualization
  • Hybrid Search
  • Media forensics
  • Multi-format Export
  • Pattern Recognition
  • Provenance tracking
  • Role-based Access Control
  • Timeline & Analysis

ArkhamMirror README

How to install ArkhamMirror

Install ArkhamMirror with Docker Compose

Docker and Docker Compose installed
Environment variables
APP_PORT, DATABASE_URL, LLM_ENDPOINT, LLM_API_KEY, VLM_ENDPOINT, EMBED_MODEL, ARKHAM_SERVE_SHELL, AUTH_SECRET_KEY, JWT_LIFETIME_SECONDS, CORS_ORIGINS, RATE_LIMIT_DEFAULT, RATE_LIMIT_UPLOAD, RATE_LIMIT_AUTH

Copy environment template

cp .env.example .env

Generate a secure auth key

python -c "import secrets; print('AUTH_SECRET_KEY=' + secrets.token_urlsafe(32))"

Add the output to your .env file

Start all services (PostgreSQL + App)

docker compose up -d

Access the application

open http://localhost:8100

Includes PostgreSQL with pgvector, all shards and UI in single container

ArkhamMirror releases & version history

More in Data Visualization

Rate this project

Click to rate — no account needed.

0 comments

Posting as guest
Loading comments…