@imageN reference not working in a Seedance prompt: ordering and content array pitfalls

When an @imageN reference fails in a Seedance prompt, the cause is usually a mismatch between the citation index in your prompt and the order of items in the metadata.content[] array, or exceeding the per request limits. RelayDance maps @image1 to the first entry, @image2 to the second, and so on, so any reordering, gap, or over limit item breaks the link. Correct the array order and stay within limits.

How the @imageN to content[] mapping works

The @imageN token in your prompt points to the Nth item in the metadata.content[] array by position, not by filename. Reference media are placed in metadata.content[] and cited in the prompt as @image1 to @imageN. If you list a video first and an image second, then write @image1, the token resolves to the wrong media type. Keep the array order aligned with your intended citation numbers. This mapping runs through the OpenAI-compatible endpoint at https://relaydance.com/v1, so the same rules apply to every request submitted with POST /v1/video/generations. See the reference details at relaydance.com/docs.

Per request limits that silently break references

A reference stops working when you exceed the per request media limits, so count each type before submitting. RelayDance allows up to 9 reference images, 3 reference videos, and 3 audio tracks per request, with clips up to 15 seconds. If you cite @image10, no tenth image slot exists and the token cannot resolve. The same applies to over long clips: a segment beyond 15 seconds will not attach. Verify these four numbers (9 images, 3 videos, 3 audio, 15 seconds) against your content[] array. Because 「Failed or errored requests are never billed」据 relaydance.com 官方文档, a rejected over limit request carries no charge, but it still returns no video, so fixing the array is the only path to a result.

Steps to fix a broken @imageN reference

Correct the reference by aligning the array order, checking limits, and reusing your existing SDK configuration.

1. Open your metadata.content[] array and confirm each item's position matches its intended @imageN number.
2. Count images, videos, and audio separately; keep them within 9, 3, and 3 respectively.
3. Confirm every clip stays within 15 seconds.
4. Set the SDK target correctly: 「Set base_url to https://relaydance.com/v1 and keep your OpenAI SDK」据 relaydance.com/docs 官方文档.
5. Add Authorization: Bearer YOUR_API_KEY using a key from the console.
6. Resubmit with POST /v1/video/generations and poll GET /v1/video/generations/{task_id} until status is succeeded or failed.

Cost context while you debug

Debugging references carries no penalty because failed requests are not billed, and successful runs bill per generated video by model rate. Current published rates include Seedance 2.0 720p at about $0.190 per second, Seedance 2.0 1080p at about $0.470 per second, Seedance Fast at about $0.152 per second, and Seedance native 4K at about $4.90 per 5-second clip. The table below summarizes these values. Live per model rates are at relaydance.com/models.

FAQ

Does @image1 refer to a filename or a position? It refers to the position in metadata.content[]. @image1 is the first array item, @image2 the second, and so on.

How many reference media can I attach? Up to 9 reference images, 3 reference videos, and 3 audio tracks per request, with clips up to 15 seconds.

Am I charged if the reference request fails? No. According to relaydance.com, 「Failed or errored requests are never billed」, so debugging your content[] array does not incur cost.