JOBE Vox
ClaimedTranscripts that know who spoke, and when. 100+ languages supported.
Paid plans only Visit WebsiteTracked since 2026
Verified vendor The Bottom Line
- Best for
- Product creators looking for long and multi-speaker audio transcription
- Price
- Paid plans only. Compare plans
Key facts
- Proven on earnings calls
- $0.08 per audio hour
- speaker diarization included
Pros
- Builtin diarization
- Crosstalk, far-field conference lines
- Jargon terminology
- Multi-speakers
- Best price in category
- 100+ languages supported
- High accuracy
Cons
- No streaming / real-time
- No word-level timestamps
- No webhooks
- No file upload
- No free tier at signup
- No SOC 2 / HIPAA
What is JOBE Vox?
JOBE Vox is an affordable, developer-friendly audio transcription API that delivers accurate transcripts with speaker diarization and timestamps at just $0.08 per audio hour.
Built for challenging recordings such as earnings calls, meetings, interviews, and conference calls, JOBE Vox identifies who spoke and when, even in long recordings with multiple speakers and noisy audio.
Integration is straightforward: provide a public or signed audio URL, submit a transcription job, and retrieve the results through a simple REST API. Vox supports MP3, WAV, M4A, and Opus files, recordings up to 2 hours long, and over 100 languages.
Originally developed to process real-world financial conference calls for StockCast, JOBE Vox combines production-tested performance with transparent, usage-based pricing.
No subscriptions, no hidden diarization fees, and no expiring credits. Pay only for the audio you process, billed by the second, with credit purchases starting at $0.50.
Available on: Web, API
Data compiled by Toolradar Pricing checked · How we evaluate
Source vox.jobe.works Preview
Key Features
- Speaker diarization
- Timestamps on every speaker turn
- Conference lines, crosstalk, financial jargon and a dozen voices
- 5.24% WER on Earnings22-Cleaned-AA benchmark
- 1.97% WER on VoxPopuli-Cleaned-AA benchmark
Pricing Plans
Pricing checked Nov 12, 2026. Source: vox.jobe.works, read Nov 12, 2026
| Plan | Price | Details |
|---|---|---|
| Pay as you go | $0.08/one-time | For everyone
|
Is JOBE Vox worth the price?
Great value
At $0.08 per hour, JOBE Vox offers exceptionally low pricing for speaker-aware transcription, far cheaper than most competing APIs.
This is a generous rate that makes it ideal for high-volume users or budget-conscious individuals and small teams.
Reviews
$99 Free with your review
Write a review Review JOBE Vox, get a free AI guide
Share your experience and we will send you Improve Your Thinking Patterns Using ChatGPT, free.
Best JOBE Vox Alternatives
Top alternatives based on features, pricing, and user needs.
ElevenLabs Freemium
Generate lifelike AI voices in 29 languages, clone voices, and enable real-time speech
Rev AI Paid
The world's most accurate API for AI- and human-generated transcripts and speech insights.
Deepgram Freemium
Enterprise Voice AI: STT, TTS & Agent APIs for accurate, realistic, and cost-effective voice solutions.
AssemblyAI Paid
Speech-to-text API with high accuracy transcription
Gladia Freemium
The speech-to-text backbone for voice agents, customer support, and meeting assistants.
Still deciding?
Most buyers shortlist 2 or 3 tools before committing. Pull a side-by-side comparison or browse the full alternatives shortlist below.
Explore More
JOBE Vox FAQ
Does JOBE Vox support speaker diarization?
Yes. JOBE Vox identifies different speakers and labels their turns throughout the transcript. When possible, it also infers speaker names from the audio. Speaker diarization is included in the standard price, with no additional charge.
How much does JOBE Vox cost?
JOBE Vox costs $0.08 per hour of audio, with speaker diarization and timestamps included.
Billing is based on audio duration, measured by the second, rather than transcription processing time.
See our benchmark results and methodology for details.
How accurate is JOBE Vox?
On the English VoxPopuli-Cleaned-AA benchmark, JOBE Vox achieved a 1.97% word error rate (WER) using verbatim mode. Lower WER indicates fewer transcription errors.
This result is based on 628 audio clips and is not an official Artificial Analysis leaderboard score. Actual accuracy depends on your audio, and the benchmark does not measure speaker diarization quality.
See our benchmark results and methodology for details.
How does the transcription API work?
Submit a publicly accessible audio URL to the JOBE Vox API. Vox checks the recording, determines its duration and price, and queues the transcription job. You can then check the job status and retrieve the transcript when processing is complete.
See the API documentation for integration details.
Do I need to upload audio files?
No. JOBE Vox works directly with audio URLs. You provide an HTTP or HTTPS URL pointing to your recording, and the service retrieves the audio for processing.
The URL must allow direct access to the file without authentication.
Which languages does JOBE Vox support?
JOBE Vox supports more than 100 languages, including English, Portuguese, and Spanish. Transcription accuracy may vary depending on the language, recording quality, and speaking conditions.
Can JOBE Vox transcribe long recordings?
Yes. JOBE Vox supports recordings up to 2 hours long. It was originally developed to process earnings conference calls, including recordings with multiple speakers, technical terminology, and challenging audio conditions.
Can I try JOBE Vox for free?
Free trial credits are available on request. Sign in and provide some information about your intended use case and expected transcription volume.
Requests are reviewed individually, and trial credits are granted at our discretion.
Does JOBE Vox store my audio recordings?
No. JOBE Vox does not store your audio files. The service access the recording through the URL you submit.
JOBE Vox does retain job information, including the submitted URL, as described in its Privacy Policy.
Is my audio used to train AI models?
No. JOBE Vox does not use your audio recordings or transcripts to train AI models.
Source: vox.jobe.works