CSV Insight Lab

cd apps/experiments/csv-insight-lab
pnpm install
pnpm dev

CSV Insight Lab parses CSV in logic.ts (max 50 data rows), builds chartReady aggregates from numeric columns, then calls generateObject for insights grounded only in those rows. The model is instructed never to invent rows. Missing AI keys return 503.

FeaturesCopy Anchor Link

• Local CSV parse – quoted fields, row cap.
• chartReady aggregates – from parsed numeric columns only.
• Insights + caveats – generateObject schema.
• Sample sales CSV – product / units / revenue.
• Honest 503 when AI is unavailable.

Server ReferenceCopy Anchor Link

POST /api/insightsCopy Anchor Link

JSON { "csv": "..." } or multipart with csv + optional file.

Success (200)Copy Anchor Link

{
  "insights": ["Units peak on Blob Drop"],
  "caveats": ["Only 5 rows"],
  "chartReady": [{ "label": "Blob Drop", "value": 200 }],
  "rowCount": 5,
  "provider": "ai-gateway"
}

Implementation DetailsCopy Anchor Link

Use CasesCopy Anchor Link

• Learn composition patterns for this lab’s stack.
• Prototype before wiring production auth and quotas.
• Deploy the standalone app as a reference implementation.

LimitationsCopy Anchor Link

• Demo-scale caps on input size and output counts.
• No auth / rate limits on public deploys.
• Model quality depends on input clarity.

Use in your projectCopy Anchor Link

Copy the Route Handler and logic.ts helpers from apps/experiments/csv-insight-lab/. Wire your own persistence and auth as needed.

DeploymentCopy Anchor Link

Deploy on Vercel

Local DevelopmentCopy Anchor Link

cd apps/experiments/csv-insight-lab
pnpm install
pnpm dev

ConfigurationCopy Anchor Link

Vercel / Next.js Features UsedCopy Anchor Link

• AI SDK generateObject
• Route Handlers
• Zod

Next StepsCopy Anchor Link