Centaur
Y Combinator 2019 · Active Healthcare Healthcare IT centaur.ai
We Create Superior Data By Making Annotation Competitive
Y Combinator's description, shown to identify the company. We rate the name, not the business.
Formerly Centaur Labs
- Evocativity EVOC : yellow
- Human-machine hybrid allusion hints at assisted AI annotation
- Brevity BREV : green
- 7 letters, 2 syllables. Up to seven letters is green, eight at two syllables too; up to ten yellow, eleven or more red.
- Greppability GREP : yellow
- Found inside 19 other dictionary words.
- Googlability GOOG : red
- Famous mythic creature has prior meaning
- Pronounceability PRON : green
- Familiar word reads naturally
- Spellability SPEL : yellow
- Unusual mythic spelling needs confirmation
- Verbability VERB : red
- No natural annotation-use action
- Nominology
- 50
- All seven dimensions and .com ownership, weighted as set out on Methodology.
- Fluency
- 100
- Brevity and pronounceability only: the subset with published outcome evidence behind it.
- PG grade
- 1 / 4
- Graham's own four-step ladder.
- Exact .com
- Not owned
- Shape
- real word
- Length
- 7 chars, 2 syllables
- Distinctiveness
- suggestive
Watkins SCRATCH flags
- Curse-of-knowledge
Rubric version 2026.10.20-1, judged by gpt-6.1-sol (agent).
This scores the name only. See Methodology.
These judged dimensions came from an AI agent working through the judging queue in a chat, on the same rubric the scoring pipeline sends. The agent passed the calibration exam before its work was accepted, and each batch it judged hid two names with known answers as a check. The model is named as the agent reported it.