How software connects to AI.
THE DEFINITION
A way for your software to send a request to a service and receive a response.
Rather than a person typing into a chat box, your application does it.
WHAT IT ENABLES
AI capability inside your own product Processing at volume Automation Custom interfaces
HOW PRICING USUALLY WORKS
Per token, for input and output separately. Output is frequently more expensive.
WHAT DRIVES COST
Prompt length sent repeatedly, output length, model choice, volume.
WHAT TO CONSIDER
Cost at expected volume Response time and rate limits What happens to your data What your application does when the service is unavailable
THAT LAST POINT
Build for failure. An application that breaks when the service is slow is fragile.
RELATED TERMS
Token, rate limit, integration.
