Exercise Videos API: Comparing GIF and Video Support

In practice, the term covers a wide range of very different products. Developer-facing REST APIs in this space typically return structured facts about an exercise (name, body part, target muscle, equipment, sometimes MET or difficulty) and either nothing visual at all, or a link to a static image or an animated GIF. Genuine video, meaning an actual video file or a streaming format like MP4 or HLS with a scrubbable timeline and optional narration, is a different product category: it shows up in paid video-course platforms and app-embedded workout players, not in most exercise metadata REST APIs. If a search result promises an "exercise videos api," read its response schema before assuming it ships MP4 files; it more often ships a GIF link or nothing visual at all.

The two formats solve different problems and cost different amounts of engineering effort. An animated GIF loops automatically, carries no audio track, and most frameworks render it with a plain <img> tag, which is trivial to drop into a card or a modal. A real video file supports a seek bar, variable playback speed, narration, and adaptive bitrate for slow connections, but needs a proper video player component and, if you want captions or multiple languages, a more complex asset pipeline. If your product only needs a short, silent, looping demonstration of a movement, a GIF is usually sufficient and cheaper to integrate. If it needs a coached, narrated walkthrough, a GIF cannot do that job regardless of which API you pick.

Neither is embedded in the metadata API. GET /v1/exercises/{id} returns facts (bodyPart, target, equipment, MET, instructions) and a HATEOAS link, _links.marketplacePreview.href, that points to /marketplace/preview/{id}.gif, a public watermarked preview of the same exercise. The clean, unwatermarked GIF is a separate purchase through the marketplace, at $0.99 per exercise with volume discounts on packs by body part, equipment or target. There is no video (MP4) file anywhere on FitExerciseDB. See the companion post on how the GIF marketplace works for pricing, packs and licence, and the guide to every field in the exercise response to see exactly what the API does return.

Coverage and rights vary a lot by source, so this is a check-per-project item rather than a safe assumption. Some community and open-source exercise datasets, the Wger project being a well-known example, publish exercise images or GIFs under a Creative Commons or similarly permissive license, which is useful for prototyping but still requires reading the exact license terms for commercial or redistribution use. Legacy scraped GIF collections that circulate without a stated source rarely come with verified commercial-redistribution rights at all, even when they are technically downloadable; treating an asset as safe to ship in a paid product just because it is publicly hosted somewhere is the most common licensing mistake in this space.

• Does the response body actually include a video or gif field, or only a separate link elsewhere?
• Is the visual asset a real video file (MP4/HLS) or an animated GIF loop, and does that match what your UI actually needs?
• Is the license scoped per project, per seat, or unlimited, and does it allow redistribution as a dataset?
• Is pricing for visual assets included in your API quota, or billed separately per asset or per pack?
• Are a free preview and a paid clean version different endpoints, and is the free preview good enough to evaluate quality before you pay?