Video pricing, in your workflow.
Query reviewed model profiles and estimate a batch. No API key needed. Generation, routing and unified billing are not enabled.
1. List supported pricing profiles
curl 'https://vidily.ai/api/market/video-models'
Use the returned exact id, profile id and durations. This reviewed list is smaller than the discovery catalog; missing normalized prices remain unknown.
2. Estimate one batch
curl 'https://vidily.ai/api/market/estimate?model=veo-3-1&profile=720p-audio&seconds=8&count=10&attempts=2&budget=100'
- model / profile
- Exact identifiers from the model response; different versions are never silently substituted.
- inputMode
- text or image. Default follows known model inputs. Unknown or mismatched endpoint modes do not receive numeric recommendations.
- seconds / count / attempts
- Supported duration; 1–10000 target clips; 1–5 planned billable attempts per clip. Minimum request charges apply before batch multiplication.
- budget
- Optional USD estimate filter, above zero and at most 1000000. This does not reserve or cap spending.
- rows / recommended
- Rows include source, checkedAt, perAttemptUsd and totalUsd. Null is unknown, never free. recommended is the lowest eligible collected quote id, or null.
Response and operating boundaries
GET only. Invalid or repeated parameters return 400. Responses declare mode: estimate_only and a calculation timestamp. Quotes expire after 14 days. Taxes, promotions, credit bonuses, deposits and provider availability are not included. There is no production SLA. Do not send secrets or prompts in query strings.
What comes next
A video gateway needs durable jobs, idempotent submissions, signed callbacks, encrypted project keys, spending controls and provider reconciliation. These are planned, not live features.