Bulk Uploading to Vecteezy: FTP Delivery and Batch Metadata
How to move from one-at-a-time uploading to batch delivery on Vecteezy — file preparation, metadata embedding and the errors that lose files silently.
# Bulk Uploading to Vecteezy: FTP Delivery and Batch Metadata
Browser uploading works until it doesn't. Somewhere between fifty and a few hundred assets, the failure modes arrive together: sessions time out, a transfer drops at 90%, the tab is accidentally closed, and you restart work you already did.
Batch delivery solves that, but only if the preparation is right. Most bulk upload problems are not transfer problems — they are file preparation problems that only become visible at scale.
## Prepare before you transfer
The rule that saves the most time: **metadata goes in before the file leaves your machine.**
Embedded IPTC metadata travels inside the file and is read on ingest. If you upload first and plan to add titles and keywords afterwards, you are editing asset by asset through a web interface — which is precisely the work batching was supposed to eliminate.
Practically:
1. Finish the assets
2. Write and embed metadata locally
3. Verify the fields actually wrote, using any metadata viewer
4. Only then transfer
Step three matters more than it looks. Metadata embedding fails quietly — an unsupported field, a tool writing to XMP but not IPTC, a format that does not carry the fields at all. Discovering that after uploading eight hundred assets is a bad afternoon.
## Transfer settings that cause silent failures
**Binary, not ASCII.** ASCII transfer corrupts binary files. They arrive, they fail validation, and often nothing tells you why. Set your client to binary for image and vector queues.
**Flat structure, not subfolders.** Most ingest processes look at a specific directory. Mirroring your local folder organisation onto the server is a common instinct and a common way to have files ignored entirely — with no error message, because from the server's perspective nothing arrived.
**Modest concurrency.** Four to eight simultaneous transfers is plenty. Dozens in parallel triggers throttling and produces more failures than the parallelism saves.
**Resume enabled.** Any interrupted transfer should resume rather than restart. For large vector files or high-resolution rasters this is the difference between a batch completing and a batch looping.
## Batch size: large enough to matter, small enough to check
There is a temptation to push everything at once. Resist it, at least until a pipeline is proven.
Several hundred assets per batch is a reasonable working size. It is enough that automation pays off, and small enough that you can verify ingest and metadata before committing the next batch. If something is systematically wrong — a metadata field not carrying, a format issue, a naming problem — you find it across three hundred files rather than eight thousand.
Once a pipeline has run cleanly several times, batch sizes can grow.
## The same assets, different metadata
Here is where multi-platform uploading goes wrong. The *assets* can usually go everywhere. The *metadata* cannot.
- Adobe Stock: up to 49 keywords, IPTC embedded, no CSV route
- Shutterstock: up to 50 keywords, categories from a fixed list, CSV applied after upload
- Getty and iStock: controlled vocabulary, no externally generated AI at all
- Freepik: mandatory `_ai_generated` flag on generative content
- Alamy: no keyword cap, AI banned by contract from 1 September 2026
- Vecteezy: free versus Pro tier decision per asset
Copying one keyword list into all of them produces rejections in some places and invisibility in others. Per-platform metadata is not a refinement; it is the thing that makes multi-agency distribution work at all.
That is the part worth automating. Rastock AI [generates per-agency metadata, embeds IPTC and delivers over FTP](/features) to 25+ destinations, so one prepared batch can go to six platforms each carrying the correct format and flags. Our [Shutterstock CSV template guide](/blog/shutterstock-csv-template) covers one platform's format in detail, and the [Vecteezy contributor guide](/blog/vecteezy-free-pro-search-ai) covers the free versus Pro decision. Plans by upload volume are on the [pricing page](/pricing).
## A workflow that holds up
1. Produce assets in batches rather than individually
2. Generate metadata for the whole batch
3. Embed IPTC and verify it wrote
4. Export per-platform metadata files where the platform uses them
5. Transfer in binary, flat structure, resumable
6. Verify ingest and metadata on the platform before starting the next batch
None of this is complicated. All of it is tedious by hand, which is why most contributors stop scaling somewhere around a few hundred files a month — not because they run out of content, but because they run out of patience for the pipeline.
This guide is part of our overview of [where to sell stock photos](https://rastock.ai/blog/where-to-sell-stock-photos), which compares every major agency on royalties, review standards and AI policy.
For the full picture, see our guide to the [stock photography workflow](https://rastock.ai/blog/stock-photography-workflow) — every stage from selection through metadata to delivery.
## Related reading
- [Bulk Metadata Generation: How to Tag Thousands of Files Without Losing Quality](https://rastock.ai/blog/bulk-metadata-generation-stock-contributors)
- [Where to Sell Stock Photos in 2026: Every Major Agency Compared](https://rastock.ai/blog/where-to-sell-stock-photos)
- [Shutterstock CSV Template: Download + How to Use It](https://rastock.ai/blog/shutterstock-csv-template)
Frequently asked questions
Can I bulk upload to Vecteezy?
Yes. Batch delivery rather than one-file-at-a-time browser uploading is the practical approach once you are submitting more than a handful of assets, and it removes the session timeouts and failed transfers that make browser uploading unreliable at volume.
Should metadata be embedded before uploading?
Yes, wherever the platform reads it. Embedded IPTC metadata travels inside the file and applies on ingest, whereas fixing metadata afterwards means editing asset by asset — which defeats the point of batching.
Why do bulk uploads fail?
Most commonly: transferring in ASCII rather than binary mode, uploading into subfolders when the platform expects a flat structure, interrupted transfers leaving partial files, and filename mismatches between assets and any accompanying metadata file.
How large should upload batches be?
Large enough to be worth automating, small enough to verify. Several hundred assets per batch is usually manageable — you can confirm ingest and metadata before committing the next batch, rather than discovering a systemic problem across thousands of files.
What file formats does Vecteezy accept?
Vecteezy's audience is vector-heavy, so editable vector formats matter alongside raster files. Supply the format buyers can actually work with, and check current platform requirements before preparing a large batch.
Can I upload the same batch to multiple platforms?
The assets, yes — but not the same metadata. Keyword limits, category systems and required flags differ by platform, so a single shared metadata set produces rejections in some places and poor discovery in others.