AI Video
AI Image
AI Music
AI Voice
Loading...
Describe the image you want to create
0/20000Work out the image before planning the motion. On Seeddance, use Nano Banana 2.1 to explore shot size, subject placement and lighting for a storyboard panel or a candidate video starting frame.
An establishing shot should explain the place; a close-up should direct attention. Write the brief around what the viewer needs to see in this particular image.
Use a required prompt of up to 20,000 characters to state camera height, shot size, subject position and visible action. Describe the moment before movement begins when preparing a starting frame.
Attach up to 14 JPEG, PNG or WebP images, each up to 30 MB. Explain which reference supplies the subject, which supplies the location and which only suggests the lighting.
Choose 16:9 for a landscape panel or 9:16 for a vertical concept when those fit your project. Check that the subject and intended direction of movement have enough room before choosing 1K, 2K or 4K output.
Nano Banana 2.1 generates and edits images. On Seeddance, treat the output as a visual proposal for a shot: a way to discuss framing, staging and lighting before making decisions about movement or timing.
This page selects Nano Banana 2.1, which is listed separately from Nano Banana 2. Fresh settings start at 1K, Auto aspect ratio and JPG. A generated panel does not establish clip length or camera movement; keep those directions in your shot notes.
Describe where a subject begins and leave appropriate space in the image. Add camera moves and action timing later in the brief for the video tool you choose.
Place downloaded panels in your own storyboard and compare eyelines, screen direction and shot size. Correct mismatches before using them to guide the next production step.
Compare whether the image communicates the intended shot, then choose the file settings for its next use.
A 1K image on Seeddance uses 2 credits; 2K uses 4 and 4K uses 6. References have no separate charge. Account for each new generation when exploring alternate framing.
For an entering character, consider open space on the entry side; for a close-up, check the eyeline and headroom. A wider ratio alone will not tell a later video model how the subject should move.
Download JPG or PNG after checking the frame. If you want to use it with a video model, manually upload the downloaded image and follow that model’s supported ratios, file limits and other input requirements.
These prompts request single still images. The review notes help decide whether each result belongs in your storyboard.
Select Text to Image and 16:9. Prompt: “One storyboard frame: a quiet roadside diner at dusk, camera across the street at eye level, lit entrance in the right third, an empty sidewalk leading in from the left, wet pavement, restrained blue and amber light, no captions.” Check that the doorway and approach path are readable.
Upload a product reference and select 9:16. Prompt: “Use the reference for the shape and color of the ceramic cup. Place the cup on the lower third of a plain table, handle facing right, soft window light from the left, open space above, static close-up, no hands and no added text.” Inspect the handle, rim and free space; describe any later hand movement separately in your video brief.
Start with the shot’s purpose and finish with an image you can evaluate in context.
Write what is visible in one frame and what the shot needs to communicate. Use Text to Image for a new setup or Image to Image with labeled subject and location references.
Choose a ratio that fits your intended display, then set 1K, 2K or 4K and JPG or PNG. Confirm the credits shown before submitting; Auto is available if you have not settled the frame shape.
Sign in and generate. Review the completed image for staging, crop and continuity with adjacent panels, then download it. Find the task again in My Creations; any later video upload is a separate manual step.
Clarify what goes into a still-image request and what needs a separate step in your production workflow.
On Seeddance, Nano Banana 2.1 costs 2 credits for each 1K image, 4 for 2K and 6 for 4K. Adding references does not create a separate reference fee. A newly generated alternative is another image request.
Yes. Use up to 14 references and describe the role of each in the prompt. Uploads support JPEG, PNG and WebP up to 30 MB per image. Public HTTP or HTTPS image links must be accessible and meet the same limits. Check the result for unintended details borrowed from either source.
There are 15 aspect-ratio options: auto, 1:1, 2:3, 3:2, 1:4, 4:1, 3:4, 4:3, 4:5, 5:4, 1:8, 8:1, 9:16, 16:9 and 21:9. Image resolution can be 1K, 2K or 4K. A later video model may accept a smaller set of ratios.
No. Nano Banana 2.1 returns a still image. Download a reviewed result, then manually upload it to your chosen video model if that model supports image input. Follow its own file, ratio and reference limits, and write the motion instructions there.
Both generation modes require a prompt, up to 20,000 characters. State the shot size, subject position and visible moment. New settings default to 1K, Auto and JPG; PNG is another output option. Verify the settings for each request. Nano Banana 2.1 has no Google Search setting.
Generation credits are refunded when the task is confirmed failed. A long wait or a temporarily unavailable status check does not confirm failure. Recheck the existing task before submitting another version of the frame.
Choose one shot, describe the staging and use Nano Banana 2.1 on Seeddance to explore its still-image composition.