AI Agents Meet AsyncAPI: Conversational Interfaces For Event Streams.
summary
Argues that AsyncAPI can let AI agents act on event streams by feeding parsed AsyncAPI definitions into an LLM and exposing small runtime functions; demo uses Confluent Kafka, Gravity catalog, KafkaJS and a Node.js agent passing function-calling via GPT-3.5-turbo to list topics, fetch schemas and read the latest message. Builds a Gravity-backed catalog, a Kafka Gateway for policying, and a working demo that returns a burgers message (type cheeseburger, pickles true, calories 147000), showing the approach is practical.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 3:16AI agents and AsyncAPI foundations
- 9:15AsyncAPI for AI agent interactions
- 13:25Event-driven dev challenges
- 14:27Demo setup and AsyncAPI catalog
- 17:35Function calling with AsyncAPI and Kafka
- 20:01Interactive demo: AI agent and Kafka
- 22:45Future possibilities and Kafka Gateway
- 24:53Future of APIs and AI agents
related talks
45mPower AI with Real-Time Data using AsyncAPI & Agent Mesh
40mASI Workshop
15mServerless & Agentic AI: Better Together
51mEvent Driven AI: Supercharging ChatGPT with RAG & LangStream
29mBeyond OpenAPI and AsyncAPI: AI agents wrote the spec, code, and tests - now w…
51mOptimizing Complex Workflows with Event-Driven Multi-Agentic Approach
40mAI Agents in Practice
50mHarnessing Event-Driven and Multi-Agent Architectures for Complex Workflows in Gen...