Reference
Command reference
aliens_eye [USERNAME ...] [options]
aliens_eye <subcommand> [options]With no usernames and no subcommand, Aliens Eye asks for everything interactively. Options set in a configuration file are overridden by the same option on the command line.
#Scanning
#What to scan
| Option | Does |
|---|---|
USERNAME ... |
One or more usernames, at least two characters each. |
-l, --level {basic,intermediate,advanced} |
Which variations of each name to try. |
--from-email ADDRESS |
Add usernames derived from an email address. |
--from-name "FIRST LAST" |
Add usernames derived from a real name. |
--from-phone NUMBER |
Add the digits of a phone number as a username. |
--profile {quick,full,aggressive} |
A preset of network settings; see Usage. |
#Which sites
| Option | Does |
|---|---|
--catalog {extended,standard} |
7,700+ sites (default) or the original 788. |
--site NAMES |
Only sites whose name contains one of these (comma-separated). |
--exclude-site NAMES |
Skip sites whose name contains one of these. |
--tag TAGS |
Only sites with these tags. |
--source SOURCES |
Only entries from these sources (maigret, sherlock). |
--no-nsfw |
Skip adult sites. |
--sites FILE |
Scan this site list instead of a catalog. |
--sites-dir DIR |
Merge the *.json site maps in this directory. |
#Detection
| Option | Does |
|---|---|
--no-ml |
Heuristics only. |
--model FILE |
Use this model instead of the shipped one. |
--no-control-check |
Skip the made-up-username check on every Found. |
--playwright |
Render Maybe pages in a browser (needs the browser extra). |
--explain |
Show the evidence behind each verdict (also with -r). |
#Network
| Option | Does |
|---|---|
-c, --concurrent N |
Requests in flight at once (default 50). |
--timeout SECONDS |
Wait for one response (default 10). |
--retries N |
Retries per failed request (default 2). |
--backoff-base SECONDS |
First retry wait (default 0.5). |
--backoff-cap SECONDS |
Longest retry wait (default 8). |
--rate-limit SECONDS |
Minimum gap between requests to one site (default 0.2). |
--max-bytes N |
Most of a page to parse (default 100000). |
--proxy URL |
http://, https://, socks4:// or socks5:// proxy. |
--tor |
Shorthand for --proxy socks5://127.0.0.1:9050. |
#Investigation
| Option | Does |
|---|---|
--correlate |
Cluster profiles that look like the same person. |
--recurse-depth N |
Scan usernames linked from found profiles, up to N hops. |
--domains |
Check <username> across common top-level domains. |
--watch INTERVAL |
Scan again every 30m, 6h, 1d, … and report changes. |
--notify URL |
POST watch-mode changes to this webhook. |
--watch-confirmations N |
Scans in a row an absence must hold before it is reported (default 2). |
--watch-profile FIELDS |
Also report changes to name, bio, avatar. |
--resume FILE |
Checkpoint to this file, and continue from it after an interruption. |
See Investigations.
#Output
| Option | Does |
|---|---|
--format FORMATS |
json, csv, html, md, pdf, gexf, mermaid, maltego, all. |
--output DIR |
Where reports go (default results). |
--only-found |
Keep only Found and Maybe in the report. |
--json-stdout |
Print the report as JSON on stdout; everything else to stderr. |
--plain |
No colours or progress display. |
-r, --read FILE |
Show a saved report instead of scanning. |
--config FILE |
Read this configuration file. |
-v, --verbose |
More detail while scanning. |
-V, --version |
Print the version. |
#Subcommands
#diff OLD NEW
Compare two saved reports. --json-stdout prints the differences as JSON; --plain
drops the colours.
#sites {stats,list,validate,import}
Inspect and import catalogs; see Sites and catalogs.
| Options | |
|---|---|
stats, list, validate |
--catalog {standard,extended} (default standard), --tag, --source, --sites FILE, --json |
import |
--maigret FILE, --sherlock FILE, --maigret-revision, --sherlock-revision, --out FILE, --with-standard, --include-examples |
#selfcheck
Measure accuracy against accounts known to exist; see Accuracy.
| Option | Does |
|---|---|
--split {train,holdout,all} |
Which ground truth to score (default train). |
--ground-truth FILE |
A custom ground-truth file instead. |
--negatives N |
Made-up usernames per site (default 1). |
--corpus DIR |
Score a frozen corpus instead of the live web. |
--allow-corpus-misses |
With --corpus, count URLs missing from it as errors. |
--detector-only |
Score the detector without the checks on every Found. |
--report {table,json} |
Output format (default table). |
--model FILE, --no-ml, --sites FILE, --plain, -v |
As for a scan. |
#corpus {record,stats}
| Options | |
|---|---|
record |
--out DIR (required), --split (default all), --ground-truth, --negatives (default 4), --plausible-ratio (default 0.5), --seed (default 1234), --concurrency (default 20), --timeout (default 15), --sites |
stats DIR |
--json |
#eval {ablate,external,groundtruth}
| Options | |
|---|---|
ablate |
--corpus DIR (required), --split (default holdout), --ground-truth, --model, --only NAMES, --fingerprints, --out FILE, --json |
external |
--corpus DIR (required), --split (default all), --ground-truth, --model, --sherlock FILE, --maigret FILE, --whatsmyname FILE, --out FILE, --json |
groundtruth |
--sherlock FILE, --whatsmyname FILE, --sites, --holdout-fraction (default 0.3), --dry-run |
#train {collect,fit}
| Options | |
|---|---|
collect |
--out FILE (default seed_dataset.csv), --negatives (default 2), --seed, --sites, --corpus DIR, --split (default train), --allow-leakage |
fit |
--data FILE (required), --out FILE (default model.json), --max-fpr (default 0.02) |
See Training the model.
#label REPORT
Label a report's uncertain results into a training CSV. --out FILE (default
labeled.csv), --band {Maybe,all} (default Maybe).
#tui [USERNAME]
The interactive terminal UI (needs the tui extra). -l/--level, --sites, --no-ml.
#serve
The MCP server over stdio (needs the serve extra). --sites, --no-ml. See
Library, MCP server and TUI.
Edit this page on GitHub
Aliens Eye 2.10.0 · MIT licence · made by Aaron Thomas