Migration wing-command to Tinyfish SDK

[KrishnaAgarwal7531]

Migrates Wing Command from raw TinyFish HTTP calls to the official SDK, and tears out the Supabase + Redis infrastructure that was never needed for this use case. The app now runs entirely in-memory — no database to provision, no cache to configure, just two API keys and you're running.
The app itself finds the best chicken wing spots near you by zip code. Pick a flavor persona, and it fires parallel browser agents at DoorDash, Uber Eats, Grubhub, and Google simultaneously, streaming results back as each source finishes rather than waiting for all of them.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 21f89981-2091-4d88-8913-0c49d786b3c2

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[simantak-dabhade]

Review: changes requested

Code migration is the cleanest pass I've seen from this batch — SDK usage is idiomatic (client.agent.stream with typed EventType / RunStatus), Supabase and Redis are fully gone from the runtime, .env.local is correctly not tracked, no shadcn UI dump. Nice work there.

Two things to fix before merge.

🔴 README is heavily out of date

The code removed Supabase + Redis + PostGIS + the Python cron scraper, but the README still documents all of it. A user following this README today would spend 30 minutes provisioning services they don't need and still not be able to run the recipe. Needs a full rewrite of the following sections:

• Architecture diagram (lines ~25-35) — currently shows Supabase (PostgreSQL + PostGIS), Upstash Redis, lib/tinyfish-scraper.ts, lib/cache.ts. None of those exist.
• Scraping Flow (line 60) — step 6 says "Cache in Redis (15-min TTL) + persist to Supabase". That path is gone; it's pure in-memory now.
• Prerequisites (line 67) — drop "Supabase project" and the Upstash Redis line.
• Environment Variables (lines 73-90) — instructs users to create .env.local with SUPABASE_SERVICE_ROLE_KEY, NEXT_PUBLIC_SUPABASE_URL, NEXT_PUBLIC_SUPABASE_ANON_KEY, UPSTASH_REDIS_REST_URL, UPSTASH_REDIS_REST_TOKEN, and TINYFISH_API_URL. The actual .env.example only needs TINYFISH_API_KEY + GROQ_API_KEY. Match them, and have README say cp .env.example .env.local instead of hand-authoring.
• Database Setup (lines 92-107) — tells users to run supabase/schema.sql. That directory doesn't exist anymore. Delete the whole section.
• Render config (lines 137-161) — render.yaml example still lists NEXT_PUBLIC_SUPABASE_URL, NEXT_PUBLIC_SUPABASE_ANON_KEY, SUPABASE_SERVICE_ROLE_KEY, UPSTASH_REDIS_REST_URL, UPSTASH_REDIS_REST_TOKEN. Strip to just TINYFISH_API_KEY + GROQ_API_KEY.
• Cron Job section (lines 166-172) — references scraper/scrape_wings.py which doesn't exist. Delete.
• Project Structure (lines 176-209) — lists lib/tinyfish-scraper.ts, lib/supabase.ts, lib/cache.ts, lib/menu.ts, supabase/, scraper/. All gone. Also mentions /api/menu/route.ts — actual routes are /api/scout + /api/discover.
• Tech Stack (lines 222-231) — drop Database: Supabase (PostgreSQL + PostGIS) and Cache: Upstash Redis.
• Constraint Checklist (lines 211-220) — still relevant but re-verify claims since infra changed.

🔴 Regression: engines field dropped

On main, wing-command/package.json has:

"engines": { "node": "22.x" }

This PR removes it. Please restore — it's one of only two recipes in the cookbook that pins Node, and the repo is moving toward standardizing on this.

🟡 Nits (not blocking)

• PR body is empty — add a short description of what was migrated.
• Next 14.2.35 — outdated vs the rest of the pack (most recipes are on Next 16). Will be handled by a separate repo-wide sweep, no action needed here.

✅ What's working well

• /api/scout/route.ts is another solid SDK reference (alongside scholarship-finder's search route).
• lib/types.ts:3 and lib/geocode.ts:3 explicitly call out "No Supabase. No Redis. Pure in-memory." — good signal that the teardown was intentional.
• .env.example format with inline sign-up URLs.
• No .env.local leak this time. (Unlike #132.)

Push README rewrite + engines restore and we'll re-review.

[simantak-dabhade]

Good progress — all 3 asks from the last review are resolved. README rewritten, engines field restored, PR body filled in. Grep sweeps all clean.

Two remaining issues before we can merge:

1. Build fails — missing isPast on CountdownTime type

npm run build fails with:

./components/CommandJumbotron.tsx:50:19
Type error: Property 'isPast' does not exist on type 'CountdownTime'.

lib/types.ts:76 defines CountdownTime with days, hours, minutes, seconds but is missing isPast: boolean. Meanwhile lib/utils.ts:113,121 returns isPast in the object, and both CommandJumbotron.tsx:50 and utils.ts:126 read it.

Fix: add isPast: boolean to the CountdownTime interface in lib/types.ts.

2. Switch from Groq to OpenAI or Gemini

The recipe uses Groq (groq-sdk, GROQ_API_KEY) for LLM calls, but Groq is currently not opening up access for new accounts. This makes the recipe impossible to run for new users — violates clone-runnable.

Please switch the LLM integration to either OpenAI (openai package, OPENAI_API_KEY) or Google Gemini (@google/generative-ai, GEMINI_API_KEY). Update:

• The SDK import and client initialization
• .env.example (swap GROQ_API_KEY for the new key, with sign-up URL comment)
• README.md env vars section and prerequisites
• package.json dependencies