Learnings from init to deploy

Summary

An agent was used to build a transcription app for a team from initialization through deployment with Stripe Projects, OpenRouter, and Vercel. The app sends video-generated subtitles through an OpenRouter agent with a detailed system prompt and Stripe documentation tool calls, producing a formatted transcription whose speaker and service names are corrected. The author recommends treating agent-generated plans like design documents, verifying provider behavior, and using streaming for long transcripts because Vercel serverless limits made non-streaming responses prone to 504s for a roughly 1,500-line transcript. Testing led to removing an unreliable free-model tier, defaulting to openrouter/auto, and choosing Claude Opus 4.5-thinking over Claude Sonnet 4.5; shared agent files were committed, and a separate review agent was used before human review.

Takeaways

1. 1

   Streaming was implemented for roughly 1,500-line transcripts after non-streaming Vercel serverless responses were judged likely to hit 10-second Hobby or 60-second Pro limits and return a 504.

2. 2

   Repeated free-model outputs condensed a 700-line script into about 100 lines, so the free tier was removed; the app defaulted to openrouter/auto, and reasoning models were found more accurate despite being slower.

3. 3

   Committing agent files such as AGENTS.md and files in the agents directory lets teammates share skills and rules, while a dedicated review agent can identify issues before the author performs the final review.