sherlock-project/sherlock

What it does

Sherlock takes one or more usernames and probes a maintained list of more than 400 social networks to see where each name appears to be registered. It is aimed at OSINT, reconnaissance, and account-handle research, and outputs found accounts to text files or optionally CSV, XLSX, or JSON. It is a name-presence checker, not an identity-matching system: it does not compare photos, avatars, or profile text.

Install

Upstream recommends pipx, pip, or uv; Docker and dnf are also listed. Third-party packages for ParrotOS and Ubuntu 24.04 are flagged as broken, so users on those systems should use uv/pipx/pip or Docker instead installation notes.

pipx install sherlock-project
sherlock user123

You can also run the container:

docker run -it --rm sherlock/sherlock

Community packages exist for Debian, Ubuntu, Homebrew, Kali, and BlackArch, but the project does not maintain or directly support them. See the installation page for all methods.

Basic usage

Pass one username or several:

sherlock user1 user2 user3

For a single username, results are written to a file named after the user, e.g. user123.txt. Multiple usernames can be sent to a folder with --folderoutput, and single-user output can be redirected with --output usage section. Useful flags include --site to limit checks to named sites, --proxy to route requests through a proxy such as socks5://127.0.0.1:1080, --csv, --xlsx, --json to load a custom site list, --timeout (default 60 seconds), --print-all or --print-found, and --nsfw to include NSFW sites. The full option list is in the CLI help.

Constraints

• Results depend on the accuracy of the site list and each network's response behavior; the CLI warns that --ignore-exclusions may return more false positives.
• The default request timeout is 60 seconds, and checking hundreds of sites can take time; use --site or --timeout to bound a run.
• NSFW sites are excluded unless --nsfw is passed.
• Requests can be sent over a proxy, but availability and blocking behavior of individual sites are outside the project's control.
• The project is MIT-licensed license.

HotRepo's take

On 2026-11-06, Sherlock ranked 20th on the Python daily trending list, adding 1,696 stars in that period, against a total of about 97.5k stars and 11.6k forks. That is a large, long-lived repository rather than a short spike, and the ~97.5k star total reflects accumulated interest since 2018. A 2019 Hacker News thread drew 172 points and 83 comments, with discussion split between practical uses such as checking whether a handle is taken comment and debate over whether username enumeration constitutes doxxing; one commenter noted the results are limited to a fixed list of sites comment. Treat those as discussion points, not product claims.