THROUGHPUTS

API Reference

Every THROUGHPUTS endpoint, parameter, and response shape.

The THROUGHPUTS API is OpenAI-compatible. If you've used the OpenAI API, you already know how to use this one — change the base URL to https://api.throughputs.dev/v1 and you're done.

Base URL

https://api.throughputs.dev/v1

All endpoints below are appended to this base.

Endpoints

MethodPathDescription
POST/chat/completionsCreate a chat completion (streaming or batch).
POST/embeddingsGenerate embeddings for input text.
GET/modelsList all models your key can access.
GET/models/{slug}Get details + pricing for a single model.
GET/pricing/plansList available plans (pay-as-you-go, coding, bundle, enterprise).
POST/pricing/estimateEstimate cost for a hypothetical workload.
GET/keysList your API keys (without secrets).
POST/keysCreate a new API key.
DELETE/keys/{id}Revoke a key.

Common request headers

HeaderRequiredDescription
AuthorizationYesBearer thp_live_xxx.
Content-TypeYes for POSTapplication/json.
X-Throughputs-Model-FallbackNoComma-separated list of fallback models. If the primary model fails or times out, THROUGHPUTS tries each fallback in order.

Common response fields

Every response includes these headers:

HeaderDescription
X-Throughputs-ModelThe model that actually served the request (may differ from model in the request after failover).
X-Throughputs-ProviderThe upstream provider (e.g. openai, anthropic).
X-Throughputs-Tokens-InPrompt tokens billed.
X-Throughputs-Tokens-OutCompletion tokens billed.
X-Throughputs-Cost-UsdTotal cost in USD with 6 decimal places.

Rate limits

Limits depend on your plan:

PlanRequests / minTokens / min
Pay-as-you-go60100,000
Coding200500,000
Bundle5002,000,000
EnterpriseCustomCustom

When you hit a limit you'll get a 429 with a Retry-After header.

Next

Read the detail docs for each endpoint:

On this page