Pricing: 5.5% Markup, No Subscription
Prepaid credits, BYOK, or usage-based billing. Pay the provider price plus 5.5%, with no monthly plan. Per-model prices are published on the models page.
Pay as you go
Add prepaid credits and pay only for what you route. Text and embedding routes are provider cost plus 5.5%. Video jobs are the direct provider quote plus 20%. Auto-refill keeps prepaid routes flowing.
Bring your own key
BYOK for any provider where you already have committed spend or enterprise rate limits. Keep your discounts; gain the attested prompt path, failover, and one unified API.
Migration credits
Teams already spending more than $100/month on LLMs can request migration credits by approval. Move your workloads over with one base_url change and keep your runway.
Cheaper. Smarter. More reliable. More secure.
5.5% prepaid markup for text and embeddings. Twenty percent for asynchronous video generation. No subscription. One API with measured routing, provider fallback, and a prompt path you can verify.
- No monthly plan or seat fees — you pay for tokens you route
- Prepaid text and embedding prices are provider cost + 5.5%, with a $0.01 per million token floor
- Video generation is the direct provider quote + 20%, billed as exact integer microdollars
- The same percentage as OpenRouter's published 5.5% pay as you go fee on credit purchases
- Per-model prices are published on the models page
- Prepaid credits, BYOK, and usage-based billing all supported
- Metadata-only billing records — no prompt or output bodies stored
- Card, ACH bank, PayPal, and stablecoin checkout for prepaid credits
# Text and embeddings bill the provider's token
# price plus 5.5%. A $0.01/M floor applies to free
# and near-zero-cost routes.
#
# Example price per 1M tokens:
#
# Provider cost $1.00
# TrustedRouter markup (5.5%) $0.055
# Customer price $1.055
#
# Video generation bills the provider's exact
# per-job quote plus 20%.
#
# Exact prompt and completion prices are listed
# on /models for every route.
#
# trustedrouter/auto picks a healthy route and
# bills at the selected provider's price.Minimum charges and per-request rounding
Eligible standard routes use the prices in effect when the request was authorized, with the rounding and minimum below. Streamed usage for these routes includes cost_microdollars.
Standard prepaid text and embedding usage is billed in whole microdollars: 1 microdollar = $0.000001. Here is how the charge is calculated:
- Start with the selected route's retail token rates and applicable context-length tier. Standard retail rates are provider cost plus 5.5%, rounded up to a whole microdollar per million tokens, with a $0.01 per million token price floor. The published retail rates already include this markup and price floor; do not add 5.5% again.
- Multiply each component's token count by its rate per million tokens, then divide by 1,000,000. Round input, output, cache read, and cache write separately to whole microdollars, half up (0.5 rounds to 1; 1.5 rounds to 2). Cached tokens use their applicable cache rates, without counting them again as ordinary input.
- Add the rounded components and any fixed per-request fee. Apply a 1 microdollar minimum per billed request whenever there is positive usage at a positive rate or a positive request fee, even if all token components round to zero. With no chargeable usage and no fee, the charge is zero.
Route-specific fees and minimums can increase the total: Perplexity routes include a per-request fee, and Parasail Liberty 2.0 has a $0.001 minimum per top-level request. Check the selected route's pricing on the models page. Video jobs use their separate quoted billing described in the video guide.
Worked examples: tiny calls
These examples use illustrative retail rates, in the same form as the published per-million prices, with no cached tokens or request fees. The raw token cost is the listed rate times tokens, before per-component rounding and the request minimum.
| Usage | Retail rates per million | Raw token cost | Rounded components | Billed | Multiple of raw cost |
|---|---|---|---|---|---|
| 10 input + 5 output tokens | $0.01 input / $0.02 output | 0.2 microdollars | 0 input + 0 output; request minimum applies | 1 microdollar ($0.000001) | 5× |
| 30 input + 10 output tokens | $0.05 input / $0.05 output | 2 microdollars | 2 input + 1 output (1.5 and 0.5 each round up) | 3 microdollars ($0.000003) | 1.5× |
On very small calls to cheap models, rounding and the request minimum can make the bill several times the raw token cost. The multiple depends on the token counts and route; the 5.5% markup alone does not explain it.
trustedrouter/auto is included on every plan.