generated: '2026-07-18'
method: searched
source: https://api.compresr.ai/api/pricing/compression-models
unit: input tokens
currency: USD
metering:
  billed_on: input tokens of the context submitted
  price_per_1m_input_tokens: 0.10
  models: [latte_v1, latte_v2]
  empty_context: no billing (empty string returns empty result)
cost_controls:
- per_key_monthly_budget: settable at key creation; exhausted budget returns 402 Payment Required
- key_expiry: optional per-key expiry date
- cost_estimate_endpoint: POST /api/pricing/estimate-cost
value_metric:
  description: >-
    Compresr is itself a FinOps lever for LLM spend — it reduces downstream model input
    tokens by 30–70% (per the SDK), so cost is measured against tokens_saved on each
    CompressionResult (original_tokens - compressed_tokens).
free_tier:
  credits_usd: 10
  card_required: false