Zandovi

Until this week, batch generation was an app feature: upload a CSV, download a ZIP. From code, rendering 200 images meant 200 POSTs, your own retry logic, and your own bookkeeping. That's done. Batch is now a first-class API resource, on every plan including Free. POST /api/v1/batch-jobs — one template, up to 400 rows, one ZIP back. What went in beyond the obvious: 🔁 Idempotent submits. Send an Idempotency-Key and a retried request returns the original job. Deduplication is a unique index in the database, not a hopeful check — two concurrent retries can't both create a job. ✅ Validation before billing. Every row is checked against the template's variable rules at submission. One bad row rejects the whole submission with a 400 that names it: no job, no charge, no half-finished batch to reconcile. 📩 Signed completion callbacks. Add a callbackUrl and we POST you a summary when the job ends — every terminal state, not just success. Each delivery carries an X-Zandovi-Signature HMAC keyed by a per-org secret that is deliberately not your API key. At-least-once, five attempts over ~15 minutes, and whether it landed is readable on the job itself. 📄 Per-row results. GET /{jobId}/items returns every submitted row, the file it produced, and the error if it has one — a failure traces back to the data that caused it without downloading anything. ↩️ Cancel refunds. Stop an active job and every row that hadn't rendered comes back to your quota. Also shipped: flat archives (images only, no manifest), 30-day result retention, and the app's Job History now lists API-submitted jobs beside CSV ones, badged with the key that sent them. Row caps are per plan — 25 on Free, up to 400 on Business. Docs: https://docs.zandovi.com/api/batch and /api/completion-callbacks