Speech Revolutions

Speech Revolutions — batch speech-to-text API

Speech-to-text that knows who said what.

Speech Revolutions has the most accurate speaker diarization on the market — less than half the errors of AssemblyAI, Deepgram and ElevenLabs.

Read the quickstart

$10 in free credit. No card required.

from speechrevolutions import SpeechRevolutions

client = SpeechRevolutions()  # reads SPEECHREVOLUTIONS_API_KEY
result = client.transcribe("audio.mp3", speaker_labels=True)

for u in result.utterances:
    print(f"{u.speaker}: {u.text}")
Outputcompleted · 2 speakers · en
  • 00:00:02Speaker A Thanks for joining. Let's start with the renewal numbers.
  • 00:00:06Speaker B Sure. Net retention came in at one-twelve for the quarter.
  • 00:00:11Speaker A And the two accounts we flagged in March?
Audio hours processed each month
60,000+
Languages, detected automatically
99+
Maximum file size
10 GB
Per audio hour, all features
$0.18

Included in every request

Everything included. No add-ons.

Diarization

The best speaker labels in the industry

Less than half the speaker errors of any other provider.

  • SPEAKER_1
  • SPEAKER_2
  • SPEAKER_1
  • SPEAKER_3

Timestamps

Word-level timing

40 ms median word-boundary error. The tightest available.

  • "Net"6.12 → 6.31
  • "retention"6.31 → 6.88
  • "came"6.90 → 7.08
  • "in"7.08 → 7.16

Languages

99+ languages, mixed freely

Detected automatically — even when speakers switch mid-sentence.

Let's ship it on Friday y luego revisamos the numbers.

en 0.0–1.4es 1.4–2.3en 2.3–2.9

Output

Six formats, one call

JSON, TXT, SRT, VTT, DOCX and PDF.

.json.txt.srt.vtt.docx.pdf

Try it

Hear it on your own audio.

Upload a file or record a sample. Live, on the production API.

Input

or

Output

Options

Speaker labels
Timestamps

Upload a file or record a sample to begin. Zephyr will detect the language, process your audio, and format the results. Your completed transcript will appear here.

Benchmarks

#1 in every category.

Head to head with AssemblyAI, Deepgram, ElevenLabs and Soniox on public datasets.

Speaker diarization

DER measures missed or misassigned speaker turns.

Dataset
AMI-SDMNotSoFar
Reading
↓ Lower is better
Methodology →
  1. SpeechRevolutions Zephyr
    9.70%AMI-SDM
    10.9%NotSoFar
  2. AssemblyAI Universal 3.5 Pro
    25.3%AMI-SDM
    23.3%NotSoFar
  3. ElevenLabs Scribe V2
    33.7%AMI-SDM
    24.4%NotSoFar
  4. Deepgram Nova-3
    30.9%AMI-SDM
    28.4%NotSoFar
  5. Soniox V5
    36.6%AMI-SDM
    32.2%NotSoFar

Benchmarks run by Speech Revolutions on public datasets, with the full method and results published. Head to head with Deepgram and AssemblyAI.

Developers

Ship it in minutes.

Official SDKs for Python, JavaScript, Go and C#.

Live progress
Server-sent events for every stage.
Signed webhooks
Your own signing secret, retried on failure.
Files up to 10 GB
Multipart uploads, handled for you.
OpenAPI 3.1
Generate a client in any language.
API reference
from speechrevolutions import SpeechRevolutions

client = SpeechRevolutions()  # reads SPEECHREVOLUTIONS_API_KEY
result = client.transcribe("audio.mp3", speaker_labels=True)

for u in result.utterances:
    print(f"{u.speaker}: {u.text}")

Install

Pricing

One price. Every feature.

Diarization, timestamps, language detection and every output format, included.

Standard Batch

High-priority processing with results delivered in seconds.

Price

$0.003/ min

Delivery

Seconds

Interactive APIs

Economy Batch

Uses idle compute for the same accuracy at a lower cost.

Price

$0.0015/ min

Delivery

≤ 24 hours

Large backlogs

Coming soon

Need volume pricing?

Volume discounts · Dedicated concurrency · Custom SLAs

See full pricing details →

Talk to our team

$10 free credit on signup

FAQ

Questions, answered.

We focus exclusively on batch transcription, with transparent per-minute pricing and no add-on fees for diarization, timestamps, or repunctuation. If you process large volumes and want fast batch turnaround at a lower cost, we're built for that.

Not at launch. Speech Revolutions is batch-only today — upload audio, track progress via SSE, download structured output. Live streaming is on the roadmap.

Standard runs on dedicated compute, and most files come back in seconds. Economy is not available yet: when it ships it will use idle capacity at ~$0.09/hr with a 24-hour delivery SLA or explicit failure notification, for large backlogs where cost matters more than speed.

No. We do not use customer audio to train speech recognition models. Uploaded files and transcripts are deleted within 30 minutes of job completion.

JSON (with word timestamps and speaker labels), TXT, SRT, VTT, DOCX, and PDF — all from a single API call.

Up to 200 MB via direct REST upload, or up to 10 GB through the SDK using presigned multipart uploads.

Start with $10 of free credit.

That's 55 hours of audio. No card required.

Talk to us