Building an AI Agent on Top of Your API
summary
Builds a working AI agent for the HiPerWare energy‑monitoring platform that connects LLMs to product APIs via LangChain, defines callable tools from an OpenAPI spec, and surfaces live data and configuration changes through Chainlit; shows code snippets and prerecorded screencasts. Describes five iterative approaches, the chosen hybrid of hand‑coded examples plus generated tooling, and concrete pitfalls: tool‑count and budget limits, fragile autogenerated prompts, nested JSON inputs, and how those problems get fixed in a deployable agent.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:50AI agent definition
- 3:37Frameworks for agents
- 7:12Hyperbot use case
- 8:37Tools and API challenges
- 12:11REST to agent tools
- 13:18Five API conversion attempts
- 16:40Manual vs automated
- 21:39Wipe coding solution
- 25:38API conversion guidelines
- 33:09Multi-agent limits
- 35:20Future of AI agents
- 36:48Key to success
- 37:09Q&A
related talks
26mAI Agents: How Do They Work and How to Build Them
40mAgents in Action: Advancing Open Source AI Missions
34mAI Agents: Transforming Autonomous Decision-Making
01h 23mBuild your own AI Agent
26mIf your APIs aren’t agent‑ready, your AI agent is just chat: [...]
01h 28mAI Coding Agents and how to code them
22mHOW TO BUILD YOUR OWN AI AGENT: THE COMPLETE ARCHITECTURE FOR ENGINEERING TEAMS
30mTutorial: Building an Observable AI Agent With an Open Source Stack