Beacon's cost is driven almost entirely by speech transcription — a few cents per child assessed. Pricing scales the same way: free to pilot, then priced per completed assessment as you grow.
A "completed assessment" is one finished ASER, EGRA, or EGMA session — however many recordings or attempts it took to get there.
For testing Beacon against real classroom audio before rolling it out.
For organizations running ongoing assessment programs. Rate depends on scheme — see below.
For large-scale survey rounds and multi-district rollouts. Rate depends on scheme — see below.
For state education departments and large NGOs needing unlimited or flat-fee licensing.
Standard and Volume prices above are "from" rates. Here's what each scheme actually costs, based on how much of it is voice vs. text.
| Scheme | Voice is required for | Standard rate | Volume rate |
|---|---|---|---|
| EGMA — Math | Arithmetic fluency only — quantity discrimination & missing-number items can be answered by text | $0.15 | $0.06 |
| ASER — Quick check | Hindi/English reading tracks are voice-only; the math track can be voice or text | $0.25 | $0.10 |
| EGRA — Full reading | All three fluency tasks — letter naming, nonsense-word decoding, oral reading | $0.40 | $0.16 |
Beacon's real cost is almost entirely the speech-to-text call behind each voice recording. EGMA's discrete math items (quantity discrimination, missing number) can be answered by text, so only its arithmetic-fluency task ever needs transcription — a fraction of a completed session. EGRA's reading-fluency tasks are always spoken, so every EGRA session carries the full transcription cost. ASER's adaptive reading tracks are voice-only too, but its math track can go either way, putting it in between.
Larger programs get a lower per-assessment rate automatically as volume grows, the same way the underlying transcription cost improves with scale.
Beacon's scoring rubric is a solid starting point, not yet a certified instrument — it's still being reviewed and tuned against real recordings. The free tier exists so your real-world audio can help validate it before you're paying for results you're relying on.
Full endpoint reference, request/response schemas, and examples for all three schemes.