Resources
Guides
Practical, implementation-level guidance for reliable file workflows. These guides describe behavior Convyx supports today, including limits and failure modes.
Batch converting files without a script
The signed-in dashboard accepts a set of files, routes each one to the correct engine queue, and keeps progress visible per file.
- Sign in, open the dashboard uploader, and add the files you need to process.
- Choose one target format that is valid for the selected sources, or select only the compatible subset.
- Review the per-file states: queued, processing, finalizing, completed, or failed.
- Download successful outputs individually or as a ZIP. A failed item can be retried without restarting the completed files.
Operational note: Batch creation is rate-limited to 240 accepted conversion requests per minute. Per-user concurrency still follows the account plan, so a large batch drains in controlled waves instead of starting every process at once.
Receiving conversion webhooks safely
Business and Enterprise accounts can register endpoints in the dashboard. Webhooks are sent by asynchronous worker jobs; synchronous /api/v1/convert responses do not generate them.
- Create an HTTPS endpoint that reads the raw request body before JSON parsing.
- Register the endpoint and keep its per-endpoint signing secret outside source control.
- Compute HMAC-SHA256 over the raw body and compare it with X-Convyx-Signature using a timing-safe comparison.
- Store X-Convyx-Delivery-Id before doing side effects so a repeated delivery is idempotent.
- Return a 2xx response quickly; move slow downstream work to your own queue.
Operational note: Canonical events are conversion.completed, conversion.failed, and conversion.expired. Delivery is retried with exponential backoff, so receivers must expect duplicates.
Choosing a retention window
Retention starts when conversion completes, not when upload begins. That prevents a slow media job from consuming its own download window while it is still processing.
- Use delete-immediately for transient workflows. It includes a five-minute download grace period rather than deleting the result before it can be retrieved.
- Use one hour for short handoffs where the recipient is ready.
- Use 24 hours for ordinary asynchronous work and timezone differences.
- Use seven days only when the plan allows it and the operational convenience outweighs keeping another copy.
Operational note: Expiration hard-deletes source and result objects from storage. A short-lived, signed download URL is generated when requested; permanent public URLs are not stored.
Lossless and lossy compression in plain English
Lossless formats preserve the represented data exactly. Lossy formats remove information to reduce size. Neither choice is universally better.
- Keep a lossless or original source for editing and archival use.
- Use lossy delivery formats for photographs, audio, or video when smaller transfers matter.
- Avoid repeatedly converting between lossy formats; each generation can discard more information.
- Evaluate output at its real viewing size and listen or inspect the details that matter to the use case.
Operational note: Format labels describe the compression model, not a guaranteed byte-size outcome. A tiny or already optimized input can become larger after conversion because container overhead and encoder settings differ.
Building an integration? Read the complete API and webhook reference.