0x6rss/instagram-private-graph

What it solves

This project provides a local research application for authorized Open Source Intelligence (OSINT) investigations on Instagram. It helps investigators analyze relationships between accounts by collecting visible information from an authenticated session and producing ranked candidate lists and network visualizations to assist in case triage and hypothesis generation.

How it works

The application uses an authorized test account to collect data from various Instagram endpoints (such as recommendations, profile context, and tagged content). It then processes this data through a local model that calculates association scores based on how often a candidate account appears across multiple discovery runs. These scores are categorized into confidence tiers (e.g., Very high, High, Medium) and visualized as a radial network graph where proximity to the target indicates the strength of the inferred association.

Who it’s for

It is designed for legitimate OSINT professionals and investigators who have authorized access to the targets they are analyzing and use dedicated, low-value test accounts to avoid account bans.

Highlights

• Relationship Inference: Uses a repeated-appearance model to estimate the likelihood of a connection between accounts.
• Network Visualization: Generates a radial graph to visualize inferred associations and their confidence levels.
• Comprehensive Reporting: Produces human-readable reports, target summaries, and technical exports in JSON, CSV, and GEXF formats.
• Local-First Security: Bound to 127.0.0.1 by default to prevent remote access and emphasizes encrypted local storage for case artifacts.