Idea.AI.

Platform starter · Shipped

// A Next.js starter for AI meeting and agent products—auth, agents, Stream video calls, tRPC APIs, Inngest jobs, and Polar billing.

role = "Full-stack engineer"

timeline = "2025"

stack = "Next.js / TypeScript / tRPC / Inngest"

impact.md

// Impact at a glance

- Agents, meetings, and call routes already wired
- Auth, billing webhooks, and background jobs included
- Designed to fork for new AI collaboration products

Next.jsTypeScripttRPCInngestStreamOpenAIDrizzlebetter-authPolar

$ open demo $ git clone $ cat details

summary.md

// summary

Idea.AI is a production-shaped starter I use to spin up AI collaboration ideas quickly: sign-in, agent CRUD, meeting rooms with Stream, background workflows, and upgrade paths wired through better-auth and Polar.

problem.md

// problem

New AI product ideas repeat the same scaffolding—auth, database, agents, realtime video, webhooks, and background jobs—before you get to the actual idea.

// what I built

I packaged the common stack into one repo: dashboard routes for agents and meetings, call pages, tRPC routers, Inngest functions, Drizzle on Neon, OpenAI + Stream integrations, and billing hooks so experiments start from a real foundation.

// core experience

- Create and manage agents, then join meetings with Stream-powered video/chat
- Typed client/server contracts via tRPC and TanStack Query
- Upgrade flow and webhooks for subscriptions without rebuilding auth each time

architecture.md

// architecture

- Next.js 15 App Router with (auth), (dashboard), and call layouts
- tRPC + Drizzle ORM + Neon; Inngest for durable background work
- better-auth, Polar SDK, Stream video/chat, OpenAI realtime helpers

// ai involvement

Agents and meeting flows are the AI surface area—OpenAI and Stream are integrated where collaboration actually happens, not as a disconnected chat widget.

challenges.md

// challenges

- Keeping the template modular enough to fork without carrying dead weight
- Wiring webhooks, Inngest, and auth without fragile local-only setups
- Making realtime and agent state understandable in the UI

// outcome

Open-source repo and live demo—ready to clone for the next AI meeting or agent product experiment.

why.md

// why this matters

Shows platform thinking: I do not only ship one-off features; I build reusable bases that compress time to a credible MVP.

reflection.md

// reflection

A starter only earns its keep if it is opinionated about the hard parts—auth, realtime, jobs—and flexible everywhere else.

capabilities.md

// capabilities

Platform architectureRealtime systemsBackground jobsAuth and billing

links.md

// links