Introduction
Tapotik AI is one platform for chat, image, voice and autonomous agents — exposed through a single, consistent REST API with typed SDKs for TypeScript, Python and Go.
Every product shares the same primitives: one API key, one billing meter (credits), one webhook system and one permissions model. You can prototype with the curl examples in these docs, then graduate to an SDK without changing any concepts.
Base URL
All API requests are made over HTTPS to the same origin. Requests over plain HTTP are rejected, as are requests without authentication.
https://api.tapotik.ai/v1v1 is stable and covered by our deprecation policy of 12 months notice.Explore the products
Chat API
Multi-model completions with streaming, tools and memory.
Image API
Production-grade image generation, styles and upscaling.
Voice API
Lifelike text-to-speech and voice cloning in 40+ languages.
Agents
Autonomous multi-step agents with tools and approval gates.
How the platform fits together
Requests are routed through our global edge network (32 regions) to the nearest inference cluster. You choose a model per request — our in-house tapotik-2 family or frontier models from OpenAI, Anthropic and Google — or let router/auto pick the best model for cost and quality automatically.
Usage is metered in credits: roughly 1,000 text tokens, one standard image, or one second of generated media per credit. Every response includes a usage object so you can reconcile spend in real time, and the dashboard breaks usage down per project, model and API key.