AirCaps AI

We've tested across laptop mics, phone mics, conference room mics - it holds up everywhere. And overlap is where you'll see the widest gap between us and others

VoiceArena independently tested A5S Async v1 on 139 natural conversations with heavy overlap and up to 8 speakers. AirCaps ranked #1 of 15 models at 23.88% cpWER, vs. 32.43% for NVIDIA Nemotron 3.5 Transcribe (released a week earlier) and 35.73% for the previous leader, ElevenLabs Scribe v2.

A5S Async v1 is also fast and cheap. 1h of audio takes 9.8s to transcribe and costs $0.18. The only model that’s faster (Deepgram) makes 3.1 times as many errors and costs >2x more per hour.

A5S Async v1 ranks #1 on 5 public multi-speaker benchmarks. Getting the words right isn't enough. You need to know who said what, even when people talk over each other. We open-sourced the benchmark harness so anyone can reproduce the results: aircaps.ai/research/a5s-asyn…

Introducing A5S Async V1: the world's most accurate speech recognition model for speaker-attributed transcription. #1 on VoiceArena, ahead of ElevenLabs, AssemblyAI, Deepgram, NVIDIA, Meta & more. The next-best model makes 41% more errors across 5 public benchmarks. And yes, every lab says they're #1. So we open-sourced our benchmark. Try it now👇

Introducing AirCaps Diarization-v1: the world's most accurate speaker diarization model. Beats every leading model across 7 popular diarization benchmarks. 61–71% fewer errors than leading cloud ASR providers with built-in diarization (ElevenLabs, AssemblyAI, Meta and Grok). 2.5x fewer than ElevenLabs Scribe v2, the best of them. 12.03% DER vs. 13.21% for DiariZen Large v2 and 13.44% for pyannoteAI Precision-3. Built for distant mics, noisy rooms and overlapping speech. Benchmarks, evaluation code & methodology open-sourced. Examples & access: research.aircaps.com/diariza…

Introducing A5S V2: SOTA streaming speech-to-text for the real world. Ranks ahead of eight leading proprietary cloud streaming speech models in a real-world accuracy evaluation derived from four public noisy, multi-talker datasets. A5S V2 achieves a 26.27% word error rate across the fixed four-corpus evaluation, beating ElevenLabs, Meta, AssemblyAI, OpenAI, xAI, Deepgram, and Google. The closest cloud model, ElevenLabs Scribe v2 Realtime, scored 26.64% WER. 36.7% lower WER than Nemotron. 47.9% lower than Whisper Streaming. Runs locally on iPhone & MacBook, with 50 ms / 44 ms mean finalization latency in our device trials. Built for noisy rooms, distant microphones, and overlapping speech. Demos, benchmarks & evaluation access: research.aircaps.com/a5sv2

Today, we’re publicly launching the AirCaps Audio Research Lab. Real-world audio is 10x harder to understand than the digital world, and voice is far from “solved.” We’re building fully streaming speech and audio models for the physical world: target speaker extraction, far-field enhancement, diarization, separation, and speech-to-text. Our models are a step function ahead of leading cloud speech models in the hardest acoustic environments, while running on edge on consumer hardware. See real demos: research.aircaps.com/