Now you generate videos with Flixly

You have 40 product photos and a launch in an hour. You open the dashboard and start the workflow at /dashboard/image-to-video.

Pick the right model first

Seedance 2.0 handles motion from stills at 24 fps for clips up to 8 seconds. Kling 3.0 adds camera moves when reference frames are supplied. Veo 3.1 works for longer narrative takes but costs 12 credits per second.

Set image-to-video parameters

Upload the first photo at 1024x1024. Choose 6-second duration, 30 fps, and motion strength 0.65. The interface shows credit cost: 48 credits total.

Add motion reference

Switch to /dashboard/reference-to-video and drop three more angles as reference frames. This keeps product orientation consistent across the sequence.

Run the first pass

Hit generate. The job finishes in 42 seconds and returns a 720p MP4. Play it back at 0:00-0:06 to check if the product rotates 45 degrees as requested.

Extend the clip

Open /dashboard/first-to-last-frame. Feed the last frame of the first clip as the new start frame. Set end frame to a different product angle. Generate a second 6-second segment.

Stitch segments

Download both MP4 files. Use the built-in editor to join them at the 6-second mark. The timeline shows total length 12 seconds.

Apply video effects

Navigate to /dashboard/video-effects. Select the combined clip and apply a subtle zoom on the product label. The effect runs at 0.3x speed between 0:03 and 0:09.

Add captions and audio

Go to /dashboard/auto-captions. Upload the 12-second file. Captions appear in 1.8 seconds with 98 percent word accuracy. Then move to /dashboard/music-generation and pick a 12-second royalty-free loop at 120 bpm.

Final export

Render at 1080p, 30 fps. The file lands at 18.4 MB. Preview once more at full resolution before exporting to your launch folder.

Verify output quality

Check frame 45 for product edge sharpness. Confirm audio levels sit at -12 dB. If any frame shows blur above threshold, regenerate only that segment with Seedance 2.0 at motion strength 0.55.

Repeat the workflow

The same steps work for the next 39 photos. Batch the uploads in groups of five to keep the dashboard responsive. Each batch averages 3 minutes 20 seconds from upload to final MP4.

Image to Video produces the core motion. Reference to Video locks product angle. First to Last Frame extends length without new uploads. Video Effects fine-tunes pacing. Auto Captions finishes the deliverable.

You now own a 12-second product video ad ready for the launch page.

Troubleshooting Common Output Issues

When motion appears jittery at the 3-second mark, lower the motion strength parameter to 0.45 and regenerate the segment using the same start frame. Blurring on reflective surfaces often stems from 1024x1024 resolution combined with 30 fps output; switch to 1280x720 at 24 fps for those frames and re-stitch. If product edges soften after the reference-to-video step, upload an additional high-contrast silhouette reference at the /dashboard/reference-to-video endpoint before the next pass.

Audio drift shows up when the 12-second royalty-free loop is imported at a mismatched sample rate. Always convert loops to 48 kHz before the auto-captions stage. Frame 45 sharpness checks can be automated by exporting a single PNG via the built-in frame grabber and comparing pixel variance against the original product photo.

Optimizing Credit Usage Across Workflows

Running image-to-video first at 6 seconds then extending via first-to-last-frame consumes 90 credits per 12-second clip. Batching five photos at once keeps the dashboard cache warm and trims average generation time by 18 seconds per item. When narrative takes require Veo 3.1, reserve it for the final 4-second hero shot only; use Seedance 2.0 for the preceding 8 seconds to stay under 60 credits total.

Monitor the live credit meter after each reference-to-video call. If the estimate exceeds 40 credits, drop one reference frame and rely on the built-in orientation lock instead. For 40-photo launches, group uploads by similar lighting angles so the model reuses motion priors and reduces per-clip credit burn by roughly 8 credits.

credit usage log helps track patterns across repeated campaigns.

Integrating with External Editing Software

Download the stitched MP4 and import it directly into an NLE timeline at 1080p. Apply color-grade LUTs only after captions are baked in, because the auto-captions layer sits on a separate track that can shift during re-encoding. When adding 3D camera moves that the internal video-effects tool cannot produce, export an alpha-matted product layer from /dashboard/video-effects and composite in the external editor.

Sync points are easiest when the 120 bpm music loop starts at the exact frame where the first motion reference ends. Export an EDL file from the dashboard timeline to preserve these markers. For teams that need frame-accurate revisions, keep the original first-to-last-frame job IDs so any regenerated segment can be swapped without redoing the full stitch.

external editor sync guide lists compatible project templates.

Checklist for Launch-Ready Video Delivery

Before moving files to the launch folder, run this sequence:

• Confirm total runtime equals 12 seconds at 30 fps
• Verify product rotation matches the 45-degree request on frames 30 and 90
• Check caption timing against spoken product name within 200 ms
• Ensure peak audio sits at -12 dB with no clipping
• Re-export a 720p proxy for quick client preview

Store the final MP4 and SRT side-by-side in the launch folder so the web player can load captions without extra requests. Repeat the same checklist for every batch of five photos to maintain consistent output across the full set of 40 assets.

Organizing Project Assets Before Launch

Before starting any batch of product photos, create a dedicated folder structure on your local drive that mirrors the dashboard job IDs. Name each subfolder after the first three characters of the photo filename plus the date stamp. This prevents filename collisions when you download multiple 12-second MP4 files in quick succession. Inside each folder keep the original still, the two generated segments, the stitched file, and the SRT caption file. A simple naming convention such as 045_2024-10-12_v1.mp4 keeps everything sortable without extra metadata tools.

When you finish a batch of five photos, run a quick checksum on the final MP4 files using the dashboard’s built-in file verifier. The verifier flags any clip whose byte size deviates more than 3 percent from the expected 18.4 MB at 1080p. Move verified files to a secondary “ready” folder so you never accidentally re-process an already approved asset. asset organization shows the exact folder template most teams reuse across campaigns.

Adjusting Parameters for Different Product Types

Not every product benefits from the same motion strength or reference count. For items with fine surface detail such as jewelry or textured fabrics, drop motion strength to 0.50 and add one extra reference frame that shows the texture under the same lighting. Reflective products like watches or glass bottles often need the 1280x720 preset at 24 fps to avoid edge shimmer that appears at 30 fps. Conversely, apparel on mannequins tolerates motion strength 0.70 because fabric folds hide minor jitter.

Create three reusable preset cards inside the dashboard: one for small hard goods, one for soft goods, and one for packaged items. Each preset stores duration, fps, resolution, and reference count so you can apply them with two clicks instead of re-entering values for every photo. After the first 40-photo campaign you will notice that 70 percent of SKUs fit the small-hard-goods preset, which trims setup time per batch from 90 seconds to under 20 seconds.

Post-Production Sync with Marketing Channels

Once the 1080p files sit in the launch folder, prepare platform-specific derivatives without touching the master timeline. Export a 720p version at 4 Mbps for paid social feeds and a 480p version at 2 Mbps for email thumbnails. Keep the original 1080p file untouched so any future platform change only requires one new export pass.

When uploading to multiple channels, align the caption SRT start time with each platform’s player offset. Most social players add a 120 ms delay; shift the SRT file forward by that amount using a free subtitle editor before upload. Store the adjusted SRT files in a subfolder named after the channel so you can locate the correct version in under ten seconds during last-minute campaign tweaks. channel sync presets contains the offset values for the five most common destinations.

Scaling the Workflow for Team Collaboration

For teams larger than two people, assign each member a distinct dashboard workspace tied to the same credit pool. Workspace isolation prevents one user’s long-running reference-to-video job from blocking another user’s image-to-video queue. Share only the job IDs of approved segments so teammates can regenerate a single clip without needing the full project context.

Use the credit history export to create a weekly burn report. The report lists credits per model and per step, revealing whether the team consistently overspends on the reference-to-video stage. If that stage accounts for more than 45 percent of total credits, reduce the default reference count from three frames to two for the next campaign. The same report also surfaces which team member finishes batches fastest, allowing the group to adopt that person’s exact parameter order.

After each campaign closes, archive the workspace but keep the job IDs in a shared spreadsheet. This lets future campaigns reload the exact reference frames that worked for similar products without re-uploading assets.