benchmarks / opsis-vs-sherlock
Opsis vs Sherlock
On a verified handle with 27 real accounts, Opsis premium found 23 at 100% precision and Sherlock found 12 at 63% precision. On usernames that don't exist, Opsis premium reported 0 false hits per run and Sherlock 5.
scores
| tool | real found | recall | precision | false hits | false hits / control | time | cost / search |
|---|---|---|---|---|---|---|---|
| Opsis premium | 23 / 27 | 85% | 100% | 0 | 0 (1 run) | 112s | 2 credits |
| Opsis | 19 / 27 | 70% | 100% | 0 | 0 (1 run) | 111s | 1 credit |
| Sherlock | 12 / 27 | 44% | 63% | 7 | 5 (4 runs) | 105s | free |
Run 2026-10-30. How this was measured →
which real accounts each found
The 27 verified accounts, split by who found them (any tier counts).
- both
- 10 chess, cracked, discord, genius, github, hackerone, minecraft, pinterest, steam, typeracer
- only Opsis
- 14 crafty, cs2, epic, faceit, fishbrain, fortnitetracker, leafly, leetcode, malibucrew, microsoft, modio, survev, twitch, twitter
- only Sherlock
- 2 tiktok, xboxgamertag
- neither
- 1 polymarket
false positives
- Opsis: 0 false hits on the verified handle; 0 per nonexistent username (1 run).
- Opsis premium: 0 false hits on the verified handle; 0 per nonexistent username (1 run).
- Sherlock: 7 false hits on the verified handle; 5 per nonexistent username (4 runs): 7cups, aniworld, awsskillsprofile, bandcamp, boardgamegeek, contently, f3, forumguns, houzz, ourdjtalk, tiktok.
premium vs standard
- Opsis premium: +5 (cracked, fishbrain, fortnitetracker, leafly, malibucrew), −1 (genius) vs standard.
other usernames (unscored)
| tool | jasondraper01 | example |
|---|---|---|
| Opsis | 8 · 100s | 281 · 54s |
| Opsis premium | 6 · 111s | 316 · 113s |
| Sherlock | 9 · 103s | 148 · 88s |
more: Opsis vs OSINT Industries · Opsis vs Maigret · OSINT Industries vs Sherlock · Sherlock vs Maigret