Aliens Eye
Find a username across 7,700+ sites.
An open-source OSINT scanner. It asks thousands of sites whether a profile exists, reads each page instead of trusting the status code, checks every hit against a username that cannot exist, and tells you how sure it is.
pip install aliens-eyeWhat it does
From one username to a map of where it is used, with the evidence for every verdict.
7,700+ sites
Community site databases merged and checked live, or the original 788-site list in under a minute.
Reads the page
Thirty signals from every response, scored by a trained model and heuristics, not a bare status code.
A second opinion on every Found
Each site is also asked for a username that cannot exist. Same page? The hit drops to Maybe.
Connects the dots
Clusters profiles by avatar, name, bio and links, follows linked usernames, checks domains.
Keeps watching
Scans again on a schedule and posts what changed to a webhook.
Fits your tools
JSON, CSV, HTML, PDF and graph exports, a Python API, an MCP server for agents, a terminal UI.
How good is it
Measured on sites the model never trained on. It finds leads, not proof: open every hit before you rely on it.
Every measurement, and what it does not show → · How a verdict is reached →
Quick start
Python 3.10 or newer, on Linux, macOS, Windows or Termux.
pip install aliens-eye
aliens_eye someone --no-nsfw # every site, about eight minutesaliens_eye someone --catalog standard # the 788-site list, under a minute
aliens_eye someone --site github,gitlab --explainaliens_eye someone --correlate --format html,pdf
aliens_eye someone --json-stdout | jq -r '.variations[].sites[] | select(.status == "Found") | .url'Documentation
From a first scan to how the model is trained and measured.
Install
pip, the optional extras, Docker, Termux, from source.
Usage
A first scan, reading the verdicts, levels, proxies, reports, scripting.
Investigations
Correlate profiles, follow links, check domains, watch for changes, resume.
Reports
Every format, the JSON report field by field, reading reports back.
Sites and catalogs
The two catalogs, how they were checked, filters, your own sites, imports.
Configuration
config.json: where it is read from and every setting.
Command reference
Every option and subcommand.
Library, MCP server and TUI
The Python API, the MCP server for agents, the terminal UI.
How detection works
Features, the two judges, site rules, the checks on every Found.
Accuracy
Measured precision and recall, the frozen corpus, ablations, other tools.
Architecture
A scan end to end, and a map of every module.
Training the model
Retrain the model, label your own data, keep the holdout clean.
Contributing to Aliens Eye
Set up, the checks that must pass, how detection changes are measured.
Changelog
Every release and what changed.
Use it responsibly
Aliens Eye only requests public profile pages. It is for legitimate OSINT research, security work and finding your own footprint. Follow the law and the sites' terms, treat reports about people as personal data, and never use it to stalk or harass anyone.