Skip to main content
The Crevio API is metered pay-per-call: AI work is billed in credits drawn from your account balance. Plain CRUD (listing products, reading orders) is cheap or free; the spend comes from AI operations — chat messages, media generation, Tasks, web research, and phone calls. Two endpoints give you full visibility: GET /usage for your current plan, balance, and spend breakdown, and GET /usage/transactions for the individual metered charges behind it.

​
Check your balance and spend

GET /usage returns your plan, remaining credit balance, and a per-operation spend breakdown for the current billing period. 
curl https://api.crevio.co/v1/usage \
  -H "Authorization: Bearer $CREVIO_API_TOKEN"
A real response: 
{
  "object": "usage",
  "plan": {
    "name": "Business",
    "status": "active",
    "is_paid": true,
    "on_trial": false,
    "trial_ends_at": null,
    "renews_at": "2026-07-01T00:00:00Z"
  },
  "credits": { "balance": 18450 },
  "spend": {
    "period_start": "2026-06-01T00:00:00Z",
    "period_end": "2026-06-30T23:59:59Z",
    "total_credits": 6210,
    "by_operation": [
      { "operation": "ai_message", "description": "AI chat messages", "credits": 3120 },
      { "operation": "image_generation", "description": "Image generation", "credits": 1840 },
      { "operation": "task_run", "description": "Task runs", "credits": 980 },
      { "operation": "web_research", "description": "Web research", "credits": 270 }
    ]
  }
}
​
plan
object
Your current plan: name, status, is_paid, on_trial, trial_ends_at, renews_at.
​
credits.balance
integer
Credits remaining on the account.
​
spend
object
Period spend: period_start, period_end, total_credits, and by_operation[] (each { operation, description, credits }).

​
See individual charges

GET /usage/transactions lists each metered charge so you can attribute spend precisely. 
curl "https://api.crevio.co/v1/usage/transactions?limit=20" \
  -H "Authorization: Bearer $CREVIO_API_TOKEN"
It returns the standard list envelope (object: "list", data, has_more) — paginate it with limit and starting_after like any other list endpoint.

​
What consumes credits

Operations that run AI or external compute are metered:
OperationWhat it covers
AI messagesChat and agent reasoning turns.
Media generationImages, video, and audio (POST /images/generate, /videos/generate, /audio/generate).
TasksEach Task run (your scheduled AI workforce).
Web researchPOST /web/research and other /web/* tools.
Phone callsAI inbound/outbound calls (POST /calls), billed per minute.
Reading and managing your own data (products, orders, customers, webhooks) is not AI-metered.

​
AI media pricing

Media generation is priced in USD per unit and converted to credits. Pick the cheapest model that meets your quality bar — the default is the most economical.

​
Images (per megapixel, unless noted)

ModelPrice
fal-ai/flux/schnell (default)$0.003 / MP
fal-ai/flux/dev$0.025 / MP
fal-ai/flux-pro/v1.1$0.04 / MP
fal-ai/recraft-v3$0.04 / image

​
Video (per second, unless noted)

ModelPrice
Kling v1.6 standard (text-to-video)$0.056 / s
Kling v1.6 pro (image-to-video)$0.098 / s
MiniMax video-01$0.50 / video
Luma Dream Machine$0.50 / video

​
Audio / TTS (per 1,000 characters)

ModelPrice
ElevenLabs multilingual-v2 (default)$0.10 / 1k chars
ElevenLabs turbo-v2.5$0.05 / 1k chars
MiniMax speech-02-hd$0.10 / 1k chars
Generated media lands in your Media Library as a file_… id you can reuse — for example in a product’s images array. Always reuse the file id rather than re-generating.
Run GET /images/models, GET /videos/models, and GET /audio/models for the live model list — new models appear there before this table is updated.

​
Next steps

Conventions

List envelopes and pagination apply to transactions too.

Errors

How balance and rate issues surface.

MCP & agents

Tasks and agent work are the main credit consumers.

API reference

Full usage and media generation schemas.