A polished prototype of an AI-powered todo list. Each task can launch an OpenAI Agents SDK workflow that gathers context, creates a plan, waits for approval, and then either completes feasible in-app work or returns concrete next steps.
npm install
cp .env.example .env.local
npm run devOpen http://localhost:3000.
OPENAI_API_KEYis required for agent workflows and task understanding.- Set
OPENAI_MODELonly when you want to override the SDK default model.
The live adapter uses structured Zod outputs, hosted web search for current details, and an approval-gated function tool before recording task completion.
npm run lint
npm run typecheck
npm test
npm run test:e2e