Pricing
How we price. No mystery form.
We publish the model and the levers. You get a range on the first call.
01 · Pilot
Fixed fee · 4–6 weeks
Fixed-fee pilot
One use case, one cohort, one metric. 4–6 weeks.
- Written charter, forward-deployed team
- Up to four weekly learning cycles
- Credited against year one
02 · Cloud or VPC
Annual fee + usage
Platform + usage
Annual platform fee, plus usage per conversation or minute.
- India-region cloud or your Azure/GCP
- Weekly Learning Cycle included
- Volume tiers up to millions a month
03 · On-prem
Annual licence
Annual licence
For institutions that must own the stack.
- Dedicated GPUs in your data centre
- Self-hosted models, learning inside your walls
- Hardware sizing included in scoping
What moves the price
Five levers.
We say which matter.
- 01VolumeTiers step down as monthly calls grow.
- 02DeploymentCloud lowest; VPC adds your infra; on-prem is licensed.
- 03LanguagesEach one adds tuning and evaluation.
- 04IntegrationsSystems connected and channels beyond voice.
- 05Your baselineBring cost per contact or RPC. We model the pilot against it.
FAQ
Pricing questions.
Do you charge per seat?
No. Production is per conversation or per minute plus a platform fee; on-prem is a licence.
Is the pilot fee credited?
Yes, against the first-year platform fee if you go to production within the agreed window.
What does the platform fee include?
The reasoning engine, Model Studio fine-tuning cycles, QA analytics and daily report, integrations maintenance, and a named delivery lead.
Do you work with PSU procurement?
Yes. We structure for tenders and empanelment, including on-prem licensing and procurement documentation.