Overview

Base URL

All public v1 requests use the same public domain. OpenAI-compatible clients point to https://api.prismaticapi.com/v1; Anthropic-compatible SDKs usually point to https://api.prismaticapi.com and append /v1/messages themselves.

What you can do

• Send OpenAI-style chat requests through POST /v1/chat/completions.
• Send OpenAI Responses requests through POST /v1/responses.
• Send Anthropic-style requests through POST /v1/messages.
• Estimate Anthropic-style input tokens through POST /v1/messages/count_tokens.
• Inspect public model IDs through GET /v1/models and GET /v1/models/{model_id}.
• Use subscription quota first, then PAYG wallet fallback when it is enabled and funded.

Why teams use it

• One API key across OpenAI-compatible and Anthropic-compatible request formats.
• Public model IDs stay stable even when internal routing changes.
• Subscription quota and PAYG checks run before generation starts.
• Streaming works across the supported text endpoints.
• Model discovery lets apps render available models dynamically.

Quick start

1. Create an API key

   Open the dashboard, go to API Keys, create a key and store it as PRISMATIC_API_KEY on your server.

2. Choose a model

   Call GET /v1/models or use the public catalog. Store only the Prismatic public model ID.

3. Send your first request

   Use Chat Completions, Responses or Messages depending on the SDK or agent you integrate.

curl https://api.prismaticapi.com/v1/chat/completions \
  -H "Authorization: Bearer $PRISMATIC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "messages": [
      { "role": "user", "content": "Write a one-line launch announcement." }
    ],
    "max_tokens": 120
  }'