6. API & Integration

• VocaladAI exposes a robust, low-latency API and SDK ecosystem that lets projects and enterprises embed voice agents wherever they engage users:

• REST & WebSocket Endpoints

  • /ask (POST): Accepts text or streamed STT packets; returns generated text + TTS audio URL or data buffer.

  • /speak (POST): Converts arbitrary text into configured project voice output, with optional SSML controls.

  • /upload_dataset (PUT): Secure admin upload for new data—whitepapers, chat logs, governance docs—triggering re-indexing.

  • /retrain_dataset (POST): Kicks off incremental RAG re-training on updated dataset shards, with version tagging.

  • /languages (GET): Returns supported language list and voice-persona metadata.

---

• SDKs & Client Libraries

  • JavaScript/TypeScript: Browser & Node modules with built-in WebSocket audio stream handling.

  • Python: Async-ready package for backend integration and on-prem pipeline embedding.

  • Go & Java (Coming Q1 ’26): For high-throughput microservices and enterprise middleware.

• Platform Adapters

  • Web3 Chat Adapters: Plug-and-play bots for Telegram, Discord, and Matrix; auto-transcode text or voice queries.

  • X Spaces Bridge: Real-time audio injection module streams Vocalad’s voice into live Spaces as co-host.

  • Web Widget: Lightweight JavaScript component for instant voice-Q&A overlays on docs, dashboards, or landing pages.

This API-first design empowers teams to drop a project-trained voice agent into any communication channel—from a new Telegram AMA to a legacy web portal—without redeploying infrastructure or managing separate services.