350 hours of Saudi full-duplex customer-service speech
Bayanat Labs (Riyadh) licenses 350 hours of Saudi full-duplex customer-service speech with task annotation and commercial AI-use rights for two-way voice models, part of Bayanat Voice’s 30,000 hours of Arabic speech.
اقرأ هذه العائلة بالعربيةWhat this dataset is for
For voice-agent teams working on customer-service interaction and evaluating the labels needed for turn-taking, corrections, and task completion.
Commercial AI-use rights included; permitted uses are specified in the agreement.
What’s included
- 350 hours of Saudi full-duplex customer-service speech
- Task annotation
- Commercial AI-use rights
What Bayanat Labs adds
- One supplier: license this collection from Bayanat Labs (Riyadh), then add custom collection, annotation, RLHF or evaluation from the same team, across nine product families.
- A team of 50,000+ expert annotators, the largest in the Middle East, including vetted native speakers, linguists and licensed domain professionals across 25+ Arabic varieties, calibrated against gold standards with adjudication and an audit trail.
- In-region hosting by default, with on-prem and private-cloud options. We build no models, so we never compete with yours.
- Methods in the open: LabelBench, the Arabic Agent Reliability Lab (methods beta) and our free datasets and tools.
Technical specification: source, format, coverage, limitations & license
Bayanat Labs licenses 350 hours of Saudi full-duplex customer-service speech with task annotation and commercial AI-use rights.
- Source
- Full-duplex customer-service speech; interaction source and channel details available on request.
- Record format / annotation
- Full-duplex customer-service speech with task annotation; coverage of overlap, interruptions and other labels is confirmed in the specification on request.
- Audio format
- Audio format, sample rate and channel layout on request
- Speakers
- Count and composition on request
- Collection year
- On request
- Language / dialect
- Saudi Arabic; request variety breakdown
- Quantity
- 350 hours
- Known limitations
- Interaction source and channel details are on request. The benchmark report is on request, not published. Permitted uses are specified in the agreement. Synthetic-voice use is not implied by recognition data rights; it needs explicit speaker terms.
- License type
- Commercial AI-use rights included; permitted uses are specified in the agreement.
- Refresh cadence
- Ask about versions and updates
- Training task
- Full-duplex voice interaction and customer-service task modeling
- Last updated
- 2 November 2026
What to check in your sample
- Speaker mix, collection dates, and recording method
- Channels, overlap, interruptions, and available task labels
- Benchmark report: benchmark name, version, evaluated model, metric, and score
We’ll send the available sample, specifications and licence wording for this collection so your team can check the fit before licensing. Need it annotated, extended or evaluated on your model? We scope a fixed-scope pilot with gold-standard QA and a quality report, usually within two weeks.
Benchmark report
Ask for the benchmark report for this collection: benchmark name and version, evaluated model, metric, and score. The report can help your team judge whether the data fits your task.
Ask for the benchmark reportCommon questions
Does the dataset include overlap and interruption labels?
Task annotation is included, but the exact coverage of overlap, interruptions, channels, and other labels needs to be checked in the product specification. Request a sample that shows the fields your model needs.
Can we review a benchmark report for this collection?
Yes, on request. Ask for the report with the benchmark name, version, evaluated model, metric, and score. That context helps your team judge how closely the test matches the customer-service interactions you need to handle.