Billing

How metering works

Money is stored as integer micro-dollars (1 USD = 1,000,000 micros), so nothing rounds badly. Each completed request is priced from the token counts the engine reports:

1cost = ceil( (prompt_tokens x price_in + completion_tokens x price_out) / 1,000,000 )2 3price_in, price_out   micro-dollars per 1M tokens, per model4+ web search          +2,000 micros per successful :online search

Example on venym/qwen-27b ($0.30 in, $1.20 out per 1M): a call with 2,000 prompt tokens and 500 completion tokens costs $0.0012.

The exact figure for every call is in Dashboard, Usage, with tokens, cost and latency per request.

Other meters

• Web search on chat: $0.0020 per search, added to the request. See web search.
• POST /v1/search: $0.0020 per call with results.
• POST /v1/scrape: $0.0010 per successful scrape.
• Failed requests are not charged for tokens that were never generated.

Credits

• You prepay. The balance goes down as you use it and never goes into debt beyond the last request in flight.
• Minimum top-up is $5.00. Pay with crypto, or card where enabled.
• New accounts get $0.25 to try the API.
• At zero, requests return 402 insufficient_credits. Nothing is queued or billed after.
• Optional auto-reload tops up when you run low, set in settings.

Checking your balance

curl https://venym.io/v1/balance -H "Authorization: Bearer $VENYM_API_KEY"

1{2  "balance_usd": 12.4831,3  "plan": "payg",4  "plan_expires_at": null,5  "key": {6    "name": "production",7    "spent_usd": 3.2104,8    "spend_limit_usd": 209  }10}

• balance_usd is the account balance, shared by all your keys.
• key.spent_usd is lifetime spend counted against that key's limit. spend_limit_usd is null when there is no limit.
• Calls to this endpoint are free and are not rate limited as chat is.

Per-key limits

Give each key a spend limit to contain a leak or a runaway loop. When it is reached, chat requests on that key fail with 402 key_spend_limit while your other keys keep working. See authentication.