Powered by the best AI
Polished doesn't lock you into a single ecosystem. We route your ad generation through a carefully orchestrated pipeline of the world's best models for text, image, video, and audio.
Video Generation
Models for animating avatars, generating b-roll, and compiling final video ads.
Veo 3.1 Pro
Cinematic video generation with native audio. Multiple tiers from flagship Pro for showpieces down to Lite for fast iterations.
Used For
Seedance 2.0
Reference-driven and image-to-video on fal — cinematic motion, native dialogue options, and tight control for multi-shot ad pipelines.
Used For
Kling 2.6
Kuaishou Kling family — Pro and Std tiers for vertical-first social cuts, turbo variants for rapid drafts, and dependable motion for paid ads.
Used For
HeyGen Avatar IV
AI talking-head avatars that act as UGC creators in any language, with natural lip-sync and brand-safe consistency.
Used For
LTX‑2 Audio-to-Video
Audio-driven video on fal — upload VO or music and generate motion that locks to the waveform for hooks and beat-synced B-roll.
Used For
Image Generation
State-of-the-art models for creating photorealistic lifestyle, product, and infographic visuals.
GPT Image 2
Hero product photography and high-fidelity ad imagery — flagship for showpiece shots, standard tier for fast iteration.
Used For
Gemini 3 Pro Image
In-context image edits and reference-driven shots — change a face, swap a background, restyle a product without losing the rest.
Used For
Flux 2.0 Pro
Stylized graphic design, abstract textures, and illustrations when we need photoreal control over weird, on-brand visuals.
Used For
Text & Reasoning
Models powering our script writing, ideation, and synthesis engines.
Claude Opus 4.7
Flagship agent reasoning — plus Sonnet for speed and Haiku for tight loops across research, scripts, and creative direction.
Used For
Gemini 3.1
Gemini 3.1 Pro and Flash — long-context synthesis, structured outputs, and speedy tool use beside Claude in the orchestration stack.
Used For
Audio & Voiceover
Best-in-class models for generating voiceovers and original music tracks.
Eleven v3
Studio-grade voiceover in any voice or accent for ads, intros, and UGC narration. Multilingual model for localized campaigns.
Used For
Eleven Music
Studio-grade AI music in any style — jingles, brand themes, and custom songs with vocals or instrumental, from a single prompt.
Used For
Suno v5
Original music and instrumentals — custom songs tuned to your brand mood, tempo, and length, ready to drop under any ad.
Used For
Research & Search
Tools that give our agents access to the live web and market data.
