Reference image
A still image you provide to a generative model as a visual anchor. In image-to-video the reference becomes the first frame; in text-to-image with reference, it guides composition or style. Crucial when you need on-model results — the same character, the same product, the same setting. The photo-to-anime flow is a worked example: a selfie becomes the reference, and the model restyles it into an anime portrait while keeping the subject recognizable.
Where this fits in GoCrazyAI
Turn a text prompt or a still image into cinematic AI video using Kling 2.5 Turbo Pro, Veo 3.1, and Sora 2 — no editing skill required.
Recent guides on reference image
Sora 4.3 TikTok video: practical replacements and GoCrazyAI workflows
After OpenAI shut down Sora (Apr 26, 2026), creators switched to image-to-video models. Compare Veo 3.1 and Kling 2.5 Turbo Pro and follow two GoCrazyAI workflows for.
AI video generator workflow: Turn one product photo into multiple high-converting vertical demos
Turn a single product photo or brief into multiple 9:16 product demos and explainers using an efficient AI video generator workflow — practical steps and prompts.
Sora 4 TikTok workflow: replace short shoots with Sora-family routes via GoCrazyAI
How to replace small TikTok/Reels shoots using Sora-family models routed through GoCrazyAI. A hands-on 6–15s vertical workflow with export settings and pitfalls.