Pond5

The Pond5 keyword generator
for footage and photos

Drop a folder of clips or photos. BatchMeta writes a title, description and up to 50 keywords for every file, then hands you Pond5’s upload CSV.

No credit card required. Or use your own API key for free.

The file Pond5 asks for

One batch produces pond5_metadata.csv — the four-column upload file, filled in and ready to attach.

pond5_metadata.csvScroll →
The columns of the Pond5 upload CSV, with one example row.
OriginalfilenameTitleDescriptionKeywords
aerial_skyline_0142.mp4Aerial drone shot flying over city skyline at sunsetSmooth aerial drone footage flying over a modern city skyline as the sun sets, golden light reflecting off glass towers.aerial, drone, city, skyline, sunset, urban, buildings, golden hour, flying, cityscape
AOriginalfilename
The exact file name you upload, extension included — Pond5 matches the row by it. Headers are matched case-insensitively, but the field itself is required.
BTitle
A clear, commercial description of the clip or photo.
CDescription
Optional in Pond5's format, but included on every row — a buyer searching footage reads this before they read keywords.
DKeywords
Comma-separated, strongest first.

What every export is held to

Pond5 doesn’t publish a bulk-CSV title or keyword limit the way Adobe or Magnific do, so these are BatchMeta’s own standards, applied to every file.

Filename columnOriginalfilename

Pond5's own header, not the "Filename" every other export here uses — one more reason to download the real file rather than retype a header by hand. Matched case-insensitively, so the exact casing doesn't matter, but the column itself is mandatory.

Title lengthUp to 100 chars

BatchMeta's shared generator caps every title at 100 characters and aims well under that — a complete commercial phrase, not a keyword list.

Keywords per file25 – 50

The same range every export gets, ordered strongest first and deduplicated case-insensitively, so a full list is that many distinct search terms.

CategoryNone

Pond5's bulk CSV has no category column to fill — there's nothing to leave blank or get wrong here.

Video, described wholeFrame-sampled

Pond5 is built around footage first, so the title and keywords come from frames sampled across the entire clip — not a guess based on the thumbnail.

How it works

1

Drop the shoot

Drag in the whole folder — JPG, PNG, WebP, MP4 or MOV. Nothing uploads: the files are read in your browser.

2

Run the batch

Every file gets a title, description and up to 50 keywords, in English, unattended from start to finish.

3

Export the CSV

Download pond5_metadata.csv and attach it wherever Pond5's uploader takes bulk metadata.

Built for the way contributors actually work

Your originals never leave your device

Full-resolution photos and videos stay in the browser. Only small downscaled preview frames are sent for analysis — there is no upload step and nothing to delete afterwards.

Video is keyworded as one clip

Frames are sampled across the whole video, so a Pond5 footage submission is described by what happens throughout the clip, not just the opening shot.

Location comes from your EXIF

GPS coordinates are read locally and passed to the AI as context, which surfaces real place names in titles and keywords instead of generic scenery terms.

The other six exports come free

One analysis also produces upload-ready CSVs for Adobe Stock, Shutterstock, Freepik, Dreamstime, Vecteezy and Getty/iStock — the same files, no re-processing per platform.

Frequently asked questions

Pond5 specifics, before your first batch.

Why is the filename column called Originalfilename?

That's Pond5's own header, not a typo — it's the one platform of the seven we export to that doesn't call it "Filename". BatchMeta's pond5_metadata.csv writes the exact header Pond5 expects, so there's nothing to rename by hand before you attach it.

What keyword or title limit does Pond5 enforce?

We haven't found one Pond5 publishes for bulk CSV metadata, and this project's own rule is to never print a platform limit we can't point to a source for — Adobe's 49, Shutterstock's 50 and Magnific's 35–100 are all cited from that platform's own documentation elsewhere on this site. Absent a published Pond5 number, the export uses BatchMeta's own standard range: titles up to 100 characters, 25 to 50 keywords per file, deduplicated and ordered strongest first.

Does this work for Pond5 footage specifically?

Yes — footage is where Pond5 is best known, and it's also where a keyword generator earns its keep most: a clip has a beginning, middle and end that a single thumbnail can't describe. Frames are sampled across the whole video before it's keyworded. A video costs 2 credits, a photo 1.

Which file does Pond5 need for bulk metadata?

A CSV with the columns Originalfilename,Title,Description,Keywords. BatchMeta exports exactly that file as pond5_metadata.csv, with UTF-8 encoding; Pond5 matches each row to a file by its Originalfilename.

Are my original photos and videos uploaded?

No. Everything is processed in your browser and only small, downscaled preview frames are sent for AI analysis. Your original files never leave your device.

What does it cost?

You start with 25 free credits and no card. After that, one-time packs start at $4.99 for 500 credits and never expire — or bring your own Gemini, OpenAI or Claude API key and pay nothing.

Keyword your next Pond5 batch

25 free credits at signup — enough for a real batch, no card required. After that, 500 credits cost $4.99 and never expire.

  • 2 credits per video
  • All 7 marketplace CSVs included
  • Credits never expire

Selling on more than one marketplace? See every export BatchMeta generates.