Lucy Restyle is not a generate
Lucy Restyle changes the look of footage you already have at up to 720p, from 4 credits (1 per second). It is an edit, not a generate or a new shot.
6 min read
Midjourney V7 I2V is 88 credits at 720p
Midjourney V7 I2V is 88 credits of silent 720p (28 at 480p). Spend it when the still is already V7 and you accept no audio.
8 min read
MiniMax H3 is silent 2K video
MiniMax H3 is 104 credits to start, silent 2K, 5–15s. Arena Elo is third-party; the Versely card has no soundtrack.
9 min read
MiniMax Speech vs Gemini Flash TTS
MiniMax Speech is 2 credits per 1,000 characters, up to 60s. Gemini 3.1 Flash TTS is 12 credits per 1,000 characters, inline tags. Length versus control.
7 min read
Pruna P-Video: 2 credits per second, 1–20s
PrunaAI P-Video is 2 credits per second of output for 1–20 second clips at 1080p. Audio is unset. A draft row, not a native-audio hero.
7 min read
Runway Gen-4.5 is silent, 200 credits
Runway Gen-4.5 is 200 credits for 5s and 400 for 10s, silent. You are buying control, not native audio.
8 min read
SAM 3 segment before background remove
SAM 3 Video Segment is 1 credit to cut a matte. Bria background remove is 1 credit. VEED is 3. Segment first when you need the key.
7 min read
Seed Audio vs Suno v5.5
Seed Audio 1.0 is 4 credits for voice-shaped reads. Suno Sounds v5.5 is 2 credits for songs. They are not substitutes.
7 min read
VEED Fabric vs Kling Avatar Pro
Fabric is 40 credits at 720p. Avatar Pro is 58 at 4K. Both need a still plus audio. Pick resolution against cost.
7 min read
VEED Lipsync is 4 credits on a video
VEED Lipsync is 4 credits on footage you already have. Fabric is 40 to animate a still. You are replacing the mouth, not building the take.
7 min read
Lay B-roll under an AI voice track
Record TTS, generate product I2V, lay it under the voice, then caption. A talking head is optional.
9 min read
Batch UGC hooks from one brief
Generate silent Turbo openings, pick keepers, run a native-audio hero, then burn one caption family.
7 min read
Burn in captions after you generate
Generate is not captioned. Transcribe, pick BASIC or DYNAMIC, burn, then spot-check names.
8 min read
Check safe area before you caption
Preview TikTok, Reels, and Shorts chrome — including YouTube’s May 2026 AI overlay — then burn captions.
9 min read
Cut jump cuts out of a generate
If a hand teleports, regenerate that beat. Do not cover an artefact jump with a caption.
7 min read
Export SRT and a burned-in MP4
Burn the short-form MP4. Export an SRT sidecar from the same transcript. Two files, two jobs.
9 min read
How to extend an 8s Veo clip
Veo extend is 7s at 720p, not Omni’s 40s chain. Extend a keeper, or generate new, or storyboard.
8 min read
Keep the SKU identical across shots
One still, then I2V or R2V. Seedance 2.0 (69cr) or Fast R2V (25cr). Do not re-prompt the bottle.
6 min read
How to label AI content on Instagram
Self-disclose the post and keep C2PA if you have it. Reels bonuses depend on that disclosure.
7 min read
How to label AI content on TikTok
Use TikTok’s built-in AI-generated content toggle before you post. Hashtags are not the label.
7 min read
How to label AI content on YouTube
Tick altered/synthetic at upload. Shorts get an overlay; long-form sits below the player.
7 min read
Lock the product still, then run I2V
The still is the SKU contract. Run I2V from that frame. Do not text-to-video a product you already photographed.
7 min read
How to pick a lipsync catalog row
Match the row to the files you have: audio-to, image-to, or video-to. Credits follow the inputs.
8 min read
Silent generate, then TTS, then captions
Kling Turbo (12cr, silent), then Gemini 3.1 Flash TTS (4cr), then burn-in. Native audio is wasted when the line must be exact.
6 min read