Higgsfield CLI run log Post: How to Make a Short AI Video with Higgsfield: Three Scenes, 14 Seconds, 59 Credits Run date: 2026-11-13. Times are KST (UTC+9) unless marked UTC. Copied from the records we kept while running the CLI for this post. Left out: account and workspace details, email, job and upload IDs (shown as and so on), result file addresses, local folder paths. == CLI version (higgsfield version) higgsfield 1.2.2, built 2026-10-16 == Balance (higgsfield account status; only the credits line was kept) start 2026-11-13 15:43:50 5364.86 credits end 2026-11-13 15:58:18 5305.86 credits spent 5364.86 - 5305.86 = 59.00 credits == Model settings $ higgsfield model get seedream_5_5 Seedream 5.5 (seedream_5_5, image) PARAM TYPE DEFAULT REQUIRED CLI aspect_ratio auto,1:1,4:3,3:4,16:9,9:16,3:2,2:3,21:9 4:3 background auto,transparent,opaque auto image_references array — use repeated --image-references (or --image) prompt string — yes resolution 1k,1.5k,2k 2k CONSTRAINTS - at most 10 image_references are allowed $ higgsfield model get kling4_0 Kling 4.0 (kling4_0, video) PARAM TYPE DEFAULT REQUIRED CLI aspect_ratio 21:9,16:9,9:16,1:1,auto 16:9 duration integer 10 end_image object|null — use --end-image image_references array — use repeated --image-references (or --image) kling_element_ids array — mode text-to-video,image-to-video,reference-to-video,video-edit text-to-video prompt string resolution 720p,1080p 1080p start_image object|null — use --start-image video_references array — use repeated --video-references (or --video) CONSTRAINTS - At most 10 images are allowed, including the first and last frames. - At most 15 images, videos, and elements are allowed in total. - Prompt is required outside image-to-video. - Text-to-video does not accept reference inputs. - Text-to-video requires a fixed aspect ratio. - Image-to-video requires images only. - Reference-to-video requires at least one reference input. - Video editing requires a video. == Estimates (higgsfield generate cost --prompt ""; no credits charged) Each "===" line: time, model and settings, prompt file (see Prompts). The line under it is the CLI output. === 2026-11-13 15:43:32 seedream_5_5 --aspect_ratio 9:16 --resolution 2k (prompt img1.txt) 3.5 credits === 2026-11-13 15:43:33 nano_banana_pro --aspect_ratio 9:16 --resolution 2k (prompt img1.txt) 2 credits === 2026-11-13 15:43:34 text2image_soul_v2 --aspect_ratio 9:16 --quality 2k (prompt img1.txt) 0.12 credits === 2026-11-13 15:43:35 kling4_0 --mode image-to-video --aspect_ratio 9:16 --resolution 720p --duration 5 (prompt vid1.txt) Error: Image-to-video requires images only. Hint: Run: hf model get kling4_0 === 2026-11-13 15:45:00 seedream_5_5 9:16 2k + image_references img1 job (prompt img3.txt) 3.5 credits === 2026-11-13 15:47:28 kling4_0 image-to-video 9:16 720p 5s start-image (prompt vid1.txt) 11.25 credits === 2026-11-13 15:47:30 kling4_0 image-to-video 9:16 720p 5s start-image (prompt vid2.txt) 11.25 credits === 2026-11-13 15:47:31 kling4_0 image-to-video 9:16 720p 5s start-image (prompt vid3.txt) 11.25 credits === 2026-11-13 15:47:32 kling_4_0_flash image-to-video 9:16 720p 5s (prompt vid1.txt) 15 credits === 2026-11-13 15:47:34 kling4_0 image-to-video 9:16 1080p 5s (prompt vid1.txt) 16.25 credits === 2026-11-13 15:53:41 kling4_0 image-to-video 9:16 720p 5s start-image (prompt vid3b.txt) 11.25 credits == Prompts (passed as --prompt "") img1.txt (scene 1 image): A lone white lighthouse on a rocky headland at blue hour just before dawn. Its beam cuts through thin sea mist, the ocean below is dark and calm, and a few faint stars remain in the sky. Vertical composition with the lighthouse in the upper third, wide shot, cinematic, cool blue palette. No people, no text, no signage, no logos. img2.txt (scene 2 image, first try): Waves roll in and break against dark wet rocks at the foot of a lighthouse headland just before sunrise. White foam and spray in the foreground, the first pale orange glow on the horizon reflected on the water. Vertical composition, low angle close to the water, cinematic. No people, no text, no signage, no logos. img2b.txt (scene 2 image, re-run): Waves roll in and break against dark wet rocks on a rocky shore just before sunrise. White foam and spray in the foreground, the first pale orange glow on the horizon reflected on the water. Vertical composition, low angle close to the water, cinematic. No people, no buildings, no text, no signage, no logos. img3.txt (scene 3 image, scene 1 image as reference): The sun rises above a calm sea horizon and warm golden light spreads across the water. The same white lighthouse stands on the rocky headland on the left side of the frame, its lamp still glowing faintly. Vertical composition, wide shot, cinematic warm palette. No people, no text, no signage, no logos. vid1.txt (scene 1 clip): The lighthouse beam sweeps slowly through the mist, the mist drifts gently, the stars fade, slow camera push-in toward the lighthouse. vid2.txt (scene 2 clip): Waves roll in and crash against the rocks, foam and spray rise and fall back, the camera holds steady low near the water, the glow on the horizon brightens slightly. vid3.txt (scene 3 clip, first try): The sun slowly rises above the horizon, golden light spreads across the gentle waves, slow camera drift to the right. vid3b.txt (scene 3 clip, re-run): The sun climbs steadily upward and lifts clear of the horizon during the shot, its golden reflection on the water stretches and brightens, gentle waves roll toward the rocks, slow camera drift to the right. == Commands images, waiting for each to finish: higgsfield generate create --prompt "" --wait --wait-timeout 9m --wait-interval 5s --json clips, submitted without waiting (the CLI prints the new job ID), then collected one by one: higgsfield generate create --prompt "" --json higgsfield generate wait --timeout 9m --interval 10s --quiet --json before and after every image and clip: higgsfield account status == Runs (one line per command, as recorded) est = estimate from generate cost; bal_before / bal_after = balance just before and after; diff = bal_before - bal_after (the charge); rc_cli / rc_dl = exit codes of the CLI and of our save script. img1 | model=seedream_5_5 | params=--aspect_ratio 9:16 --resolution 2k | prompt=img1.txt | est=3.5 | start=2026-11-13 15:44:05 | end=2026-11-13 15:44:39 | elapsed_s=34.4 | bal_before=5364.86 | bal_after=5361.36 | diff=3.5 | rc_cli=0 rc_dl=0 img2 | model=seedream_5_5 | params=--aspect_ratio 9:16 --resolution 2k | prompt=img2.txt | est=3.5 | start=2026-11-13 15:45:02 | end=2026-11-13 15:45:30 | elapsed_s=28.2 | bal_before=5361.36 | bal_after=5357.86 | diff=3.5 | rc_cli=0 rc_dl=0 img2b | model=seedream_5_5 | params=--aspect_ratio 9:16 --resolution 2k | prompt=img2b.txt | est=3.5 | start=2026-11-13 15:45:49 | end=2026-11-13 15:46:25 | elapsed_s=35.4 | bal_before=5357.86 | bal_after=5354.36 | diff=3.5 | rc_cli=0 rc_dl=0 img3 | model=seedream_5_5 | params=--aspect_ratio 9:16 --resolution 2k --image-references | prompt=img3.txt | est=3.5 | start=2026-11-13 15:46:26 | end=2026-11-13 15:47:01 | elapsed_s=34.3 | bal_before=5354.36 | bal_after=5350.86 | diff=3.5 | rc_cli=0 rc_dl=0 vid1 | submit | model=kling4_0 | params=--mode image-to-video --aspect_ratio 9:16 --resolution 720p --duration 5 --start-image | prompt=vid1.txt | est=11.25 | submit=2026-11-13 15:47:54 | bal_before=5350.86 | bal_after=5339.61 | rc_cli=0 vid2 | submit | model=kling4_0 | params=--mode image-to-video --aspect_ratio 9:16 --resolution 720p --duration 5 --start-image | prompt=vid2.txt | est=11.25 | submit=2026-11-13 15:47:57 | bal_before=5339.61 | bal_after=5328.36 | rc_cli=0 vid3 | submit | model=kling4_0 | params=--mode image-to-video --aspect_ratio 9:16 --resolution 720p --duration 5 --start-image | prompt=vid3.txt | est=11.25 | submit=2026-11-13 15:48:00 | bal_before=5328.36 | bal_after=5317.11 | rc_cli=0 vid1 | wait job= | wait_start=2026-11-13 15:48:19 | wait_end=2026-11-13 15:52:23 | rc_cli=0 rc_dl=0 vid2 | wait job= | wait_start=2026-11-13 15:52:29 | wait_end=2026-11-13 15:52:33 | rc_cli=0 rc_dl=0 vid3 | wait job= | wait_start=2026-11-13 15:52:33 | wait_end=2026-11-13 15:52:37 | rc_cli=0 rc_dl=0 vid3b | submit | model=kling4_0 | params=--mode image-to-video --aspect_ratio 9:16 --resolution 720p --duration 5 --start-image | prompt=vid3b.txt | est=11.25 | submit=2026-11-13 15:53:43 | bal_before=5317.11 | bal_after=5305.86 | rc_cli=0 vid3b | wait job= | wait_start=2026-11-13 15:53:45 | wait_end=2026-11-13 15:57:27 | rc_cli=0 rc_dl=0 stitch | start=2026-11-13 15:57:59 | end=2026-11-13 15:58:02 == Credit transactions (higgsfield account transactions; times as the CLI prints them, in UTC) listed at 2026-11-13 15:58:19 KST DATE MODEL CREDITS ACTION 2026-11-13 06:53 Kling 4.0 -11.25 spend 2026-11-13 06:48 Kling 4.0 -11.25 spend 2026-11-13 06:47 Kling 4.0 -11.25 spend 2026-11-13 06:47 Kling 4.0 -11.25 spend 2026-11-13 06:46 Seedream 5.5 -3.5 spend 2026-11-13 06:45 Seedream 5.5 -3.5 spend 2026-11-13 06:45 Seedream 5.5 -3.5 spend 2026-11-13 06:44 Seedream 5.5 -3.5 spend 8 rows from this run, sum -59.00. Earlier rows in the list are other work from before this run and are left out. == Re-runs img2 -> img2b: the first scene 2 image showed two lighthouses on the headland; we took the lighthouse out of the prompt and added "no buildings". vid3 -> vid3b: in the first scene 3 clip the sun stayed at the same height; we re-ran it from the same start image with a more direct motion prompt. == Totals images: 4 x 3.5 = 14.00 (img2 not used) clips: 4 x 11.25 = 45.00 (vid3 not used) total: 59.00 = balance 5364.86 - 5305.86 in the finished video: 3 x 3.5 + 3 x 11.25 = 44.25; the two re-runs: 14.75 first generation 15:44:05 -> finished file 15:58:02 (join with ffmpeg 15:57:59 -> 15:58:02)