Pricing

Simple, honest pricing.

Pay per audio-hour processed. Batch tier only for MVP — real-time streaming coming in Phase 2. Exact rates are revealed at launch;get early access to be first in line.

Launching soon
Batch tier · Guaranteed 30-min SLA
₹10/ audio-hour(~16 paise / min)

Upload audio, get a transcript within 30 minutes. P50 delivery ~15–20 min. Up to 78% cheaper than other enterprise Indic ASR providers.

Available in every request
  • 24 Indic languages (19 tested + 5 experimental)
  • Speaker diarization
  • Indic PII detection + redaction
  • Segment-level timestamps per turn
  • SRT / VTT / JSON / plain text outputs
  • Multi-channel stereo transcription
  • Signed webhooks + polling API
  • Idempotency keys
  • Per-key + org-wide usage caps
Real-time
/ audio-hour
target price · Phase 2 · revealed at launch

Streaming ASR for live transcription, IVR, and contact-center use cases. Ships once demand supports always-on infrastructure.

Talk to us about live workloads
FAQ

Pricing questions.

How is a minute of audio priced?

You're billed on the audio duration we successfully transcribed, rounded to the nearest second. Silence, retries, and failed jobs don't count. Rates shown are per hour of audio for readability.

What currencies do you bill in?

INR and USD, both billed via Razorpay (UPI, cards, netbanking). One payment provider for every customer, regardless of billing address.

Is there a free tier?

No free tier. There is a small credit balance on signup so you can try the API end-to-end without committing. After that it's strict pay-as-you-go.

Do you charge for failed jobs?

No. Jobs that error out before writing a transcript don't generate a usage event. You only pay for what we deliver.

How does the usage cap work?

Each API key can carry a monthly audio-seconds cap and/or a monthly cost cap (₹ or $). Your organisation carries an org-wide cap on top. Whichever trips first blocks new jobs from that key. Caps reset on the calendar-month boundary.

When is real-time streaming available?

Planned for a later phase, once demand justifies keeping infrastructure always-on. Reach out if you have a live-transcription workload we should know about.

Be first in line at launch.

Get early access — we'll reach out when asr-core opens up.