v1.2.0
Reliability & Recovery
Highlights
- • Thumbnail extraction — extract still frames at specific video timestamps
- • Download URL regeneration — get fresh download links without reprocessing
- • Bulk job management — cancel or retry up to 50 jobs at once
- • Real-time usage log with live auto-refresh
- • Idempotency keys prevent duplicate jobs on client retries
- • Zombie job watchdog auto-recovers stuck processing jobs
- • Health check endpoint for uptime monitoring
Dashboard
- + Invoice history — view recent invoices, receipts, and payment status directly on the billing page
- + Billing portal — manage payment methods, download invoices, and update subscription via Stripe's secure portal
- + Download URL regeneration — 'Regenerate Download Link' button on completed jobs for fresh download links (free, no credit charge)
- + Bulk job management — select multiple jobs and cancel or retry in bulk (up to 50 at a time)
- + Real-time usage log — see individual API requests as they happen with 10-second live refresh
- + Filter requests by endpoint, status, date range, or URL search
- + Real-time stats banner showing hourly request count, credit usage, and active jobs
Developer Experience
- + Comment reply threading — /v1/comments now returns nested reply threads with reply_count and include_replies toggle
- + POST /v1/thumbnail — extract still frames at up to 20 specific timestamps per request (jpg/png/webp)
- + Custom webhook headers — add up to 10 custom headers (e.g., Authorization) to webhook deliveries
- + POST /v1/jobs/{id}/resign — regenerate expired download URLs for completed jobs at no cost
- + Idempotency-Key header support on all POST endpoints — safely retry requests without duplicate charges
- + X-Idempotency-Replayed response header indicates cached responses
- + GET /api/health endpoint for load balancer and uptime monitoring
Reliability
- + Zombie job watchdog detects and recovers stuck jobs with credit refunds
- + Stale pending job recovery re-enqueues jobs stuck for over 1 hour
- + Database connection pooling for improved throughput under load
- + Proactive disk space monitoring prevents storage-related failures
- + Failed message recovery via dead letter queue ensures no jobs are silently lost