Newsletter
Join the Community
Subscribe to our newsletter for the latest news and updates
Image, video, audio and chat APIs with prepaid USD usage billing, documented jobs, polling, webhooks and OpenAI-compatible chat.
SpicyAPI is an AI API platform for developers building applications with image, video, audio and language models. Its public documentation describes request schemas, authentication, USD price quotes, asynchronous media jobs, polling, webhooks and OpenAI-compatible chat.
Developers can connect supported models to their own applications through documented API requests. Request schemas describe the fields accepted by each model so an integration can prepare inputs before starting a job. For asynchronous media work, applications can read task status or handle webhook notifications instead of holding a browser page open until generation finishes.
The platform provides public SDKs, a CLI, a local MCP client and an agent skill. These are different ways to connect to the same hosted APIs: an SDK is intended for application code, the CLI supports terminal workflows, and the MCP client connects compatible assistants using the user's own API key.
Start with the official documentation and the supported model's request schema. Check the applicable USD price information before creating paid work. Submit the request through the documented interface, track asynchronous jobs with polling or webhooks, and use the completed result in your application. Choose the client that fits your environment rather than assuming every integration needs an MCP server.
Hosted inference is a paid service. It uses a prepaid USD balance and is billed by usage, without a recurring subscription. Current model prices and top-up information are maintained on the official pricing page.
The public client projects are MIT-licensed and free to use. Their source-code license is separate from the paid hosted service; installing a client does not make model inference free.
