A media library that knows what is in it

When you upload a file to SPREVA, it is described, tagged and, if it is a video, transcribed, before you have named it anything. Search then works on what a file shows or says, not on what it is called. The same descriptions are what let an agent pick the right picture for each line of a script, or suggest alt text before you ask.

A folder is a decision made once, at upload, about the one way you will look for this file later. “Menu / September” is fine until you need every photo with a person in it, or every clip where someone says the word “roast”, or the picture of the window at dusk that you know exists. Nobody tags for those questions in advance, and after a few hundred files nobody looks through the tiles either.

The cost is not lost files. It is the photo you reuse for the fourth time because it was on the first screen, and the good one from March that no post ever used.

Each upload is looked at as it arrives, and the result is kept beside it:

• A description of what the picture shows, in plain words: the place, what is happening, the light.
• Tags, from the same look, so “outdoor”, “people” or “product” find it later.
• A transcript for a video, so what is said in it can be searched, and clips can be cut on whole sentences.
• Alt text, suggested from the description and ready to be edited, so a post can carry it without anyone writing it at 11pm.

Search reads the descriptions and the transcripts. Type what you remember (“window at dusk”, “the clip where she laughs”, “anything with the red bag”) and the matches are the files that show or say it, whatever they are called and whichever folder they went into.

Folders still exist for what they are good at, which is keeping a client's or a campaign's files apart. They are no longer the only way in.

An agent can only use what it can find. With descriptions, “the five photos from yesterday's tasting” is a search, not a guess, and each line of a script can be matched to the picture that shows what it says. That is how a video from photos gets built in one step:

The agent finds Tuesday's photos, leaves out the ones whose descriptions mention people, writes the script and puts each line on the photo that fits it. You correct a choice if you want to: there is a picture to correct, rather than a blank. The whole flow, step by step.

Alt text is the thing everyone means to write and few posts carry. Because a description already exists, SPREVA suggests alt text from it for every image in a post, and you edit it rather than write it.

Upload a folder you have stopped looking through and search it for something you remember but never tagged. Then ask your agent for a video from what it finds. Every plan starts with a 7-day trial, and uploads take JPEG, PNG, WebP, AVIF, GIF and video up to 1 GB.