Why Compare STT exists

Every speech-to-text provider claims to be the best. They publish benchmarks on cherry-picked datasets, compare against last year's models, and sprinkle asterisks everywhere. We've all read those blog posts. We've all squinted at those WER tables. And we've all wondered: but how do they actually compare on my audio?

Compare STT was born from that frustration. Instead of trusting self-reported numbers, we let you be the judge. Record your voice, upload a meeting snippet, throw in a podcast clip — whatever you want. Two providers transcribe it blindly, side by side, and you pick the one that got it right. No names attached. No bias. Just text.

Every vote feeds into an ELO rating system — the same kind used to rank chess players. Over time, the leaderboard converges toward something no corporate benchmark can replicate: a ranking shaped by real people, real accents, real noise, and real conversations.

How it works

1. You submit audio — record live or upload a file (up to 2 minutes).
2. Two random providers transcribe it — their identities are hidden. You see “Model A” vs “Model B”.
3. You vote — pick the better transcription, or call it a tie. The providers are revealed after your vote.
4. ELO ratings update — every vote shifts the leaderboard. The more votes, the more accurate the ranking.

Current providers

Gladia

Solaria

ElevenLabs

Scribe v2

Deepgram

Nova 3

AssemblyAI

Universal-Streaming

Speechmatics

Enhanced

Mistral

Voxtral Mini

What we don't do

• We don't store your audio or transcriptions.
• We don't track you — no cookies, no accounts, no IP logging.
• We don't sell data. There is no data to sell.
• We don't take money from providers to rig the leaderboard.

The only thing we keep is the vote outcome: who won, who lost, or whether it was a tie. That's it.

Sponsored by Gladia

Compare STT is a free, open, non-commercial project sponsored by Gladia. Yes, Gladia is also one of the providers in the arena. No, they don't get special treatment. The matchmaking is random, the voting is blind, and the leaderboard is driven purely by community votes. If Gladia's model is good, it'll climb. If it's not, it won't. That's the whole point.

Want to add a provider, report a bug, or just say hi? Reach out.