Sherlock vs Blackbird: Which Username OSINT Tool Fits You?

Sherlock and Blackbird are two of the most popular free command-line tools for finding accounts by username. Both are written in Python, both take a handle and return the sites where it exists, and both are free. Under the hood they differ in where their site lists come from, what else they search, and how actively they are maintained. This comparison uses each project's README and repository data as of October 14, 2026.

Trademark note: Sherlock and Blackbird are the names of their respective projects. WhatsMyName Apps is independent and not affiliated with either. Credit for both tools goes to their maintainers and contributors.

The short answer

• Pick Sherlock if you want a well-maintained, MIT-licensed tool that is easy to install, checks several usernames per run and slots into scripts.
• Pick Blackbird if you also want to search by email address, like the idea of a site list built on the WhatsMyName dataset, or want a ready-made PDF report.

At a glance (checked October 14, 2026)

	Sherlock	Blackbird
Repository	sherlock-project/sherlock	antoniaci/blackbird (the old p1ngul1n0 address redirects here)
Searches by	Username	Username or email
Site list	Its own data file: README says 400+; the file had 481 entries when we counted	WhatsMyName dataset: README says 600+; the dataset had 717 entries when we counted
Exports	TXT (default), CSV, XLSX	PDF, CSV
Several usernames per run	Yes	Not documented in the README
Extras	Proxy, per-site runs, username variants with {?}, open hits in browser	Optional AI profiling (free API key, daily limit)
Install	pipx install sherlock-project, Docker, dnf, community packages	Clone the repo, pip install -r requirements.txt
License	MIT	No license file in the repository
GitHub stars	94,116	9,129
Latest activity	Release v0.16.2 on Sep 8, 2026; commits this week	Last commit Jul 13, 2025

Sherlock in brief

Sherlock describes itself as a way to "hunt down social media accounts by username across 400+ social networks". It was created by Siddharth Dushantha and is maintained by the Sherlock Project under the MIT license. You run sherlock user123, or several names at once, and it writes the hits to a text file per username. Flags add CSV or Excel output, limit the run to named sites, route traffic through a proxy, change the timeout, include adult sites with --nsfw, or try separator variants of a handle with {?}.

It doesn't parse profiles or build reports. It answers one question, where does this handle exist, and does it predictably. The project ships official pipx and Docker installs and warns that some third-party distro packages (ParrotOS, Ubuntu 24.04) appear broken. More in our Sherlock guide.

Blackbird in brief

Blackbird is developed by Lucas Antoniaci. Its README describes "fast username and email searches across more than 600 platforms", built on the community WhatsMyName project, and credits that dataset for low false-positive rates. You run python blackbird.py --username johndoe or --email with an address, and add --pdf for a report.

The newer feature is AI profiling. After generating a key with --setup-ai, the --ai flag returns a "behavioral and technical profile" based on the sites where the name was found. The README says only site names are sent, usage is free within a daily limit, and results go into the PDF. The README also carries a disclaimer that the program is "for Educational purpose ONLY" and should not be used without permission. Our Blackbird overview has setup notes.

The differences that matter

Where the site list comes from

Sherlock maintains its own list, with detection rules per site and a list of upstream exclusions for sites known to give false positives (you can override them with --ignore-exclusions). Blackbird reads the WhatsMyName dataset, which describes each site with both a "found" and a "not found" signature and is maintained by a separate community. Neither approach is perfect. Our own study of 3,166 sites found that sites behind bot-protection pages are a common source of false hits for any checker.

Email search

Sherlock does usernames only, so this is where Blackbird pulls ahead. If your starting point is an email address, Blackbird (or an email-first tool such as Holehe) covers a step Sherlock can't.

Reports

Blackbird's PDF is ready to hand to someone. Sherlock's TXT, CSV and XLSX files are better for further processing in a spreadsheet or a script.

Maintenance and license

Sherlock is very active, with a September 2026 release and commits in the week we checked. Blackbird's last commit was in July 2025, though its site list partly stays current because it draws on WhatsMyName data. Sherlock's MIT license makes reuse straightforward. We found no license file in Blackbird's repository, which leaves reuse of its code unclear, so read the repository before building on it.

AI features

Blackbird's AI profile is a convenience, but treat it with care. A summary of "who this person is" built from a list of site names is a guess, and it can make weak matches look like facts. Verify each account by hand before drawing conclusions.

Which one should you use?

• Scripted checks, several handles, a clear license: Sherlock.
• Started from an email address: Blackbird.
• Need a PDF for a non-technical reader: Blackbird.
• Want the deepest free option with profile parsing and recursive search: neither; look at Maigret. Our Sherlock vs Maigret vs WhatsMyName comparison covers it.
• No terminal at all: a browser tool. See the roundup of the best free OSINT username tools.

Running both is also reasonable. Their lists overlap only in part, so a site found by one and missed by the other is worth a closer look.

A browser option

If you'd rather not install Python, WhatsMyName Apps runs in the browser and checks 3,165 sites by merging the WhatsMyName, Sherlock and Maigret lists. It shows Found, Not found, Blocked and Unknown separately, adds public display names and avatars where it can, and exports JSON or CSV. It has no email search, no PDF report and no API, so Blackbird and Sherlock remain the better choices for those jobs. Try a username; the name is used only to run the check and is not stored or logged.