Use reference images instead of style prompts for stabler AI image generation
dotey · x · 2026-10-02
Baoyu (@dotey) shared a prompting tip from testing Manus image generation: give a reference image instead of describing the style in words.
- Pre-bake style elements — colors, font samples, rice-paper texture, ink edges, water marks, distant mountains, seals, and whitespace rules — into the reference image; output stability improves noticeably and the content prompt can stay minimal.
- He includes two full contentprompt examples (tea-themed minimal ink-wash slides) plus a reusable English prompt template covering hierarchy & layout guidance, visual direction, and readability/bleed constraints.
Key takeaway: let the reference image carry the style, and let the prompt handle only content and layout rules — a practical combo for consistent AI-generated visuals.
More from Multimodal
- Ben Affleck fine-tunes open video models on own-shot data; Netflix bought his AI firm for $587M — rohanpaul_ai · 2026-10-02
- ComfyUI node pack adds LLM prompt enhancer for Qwen, MiniMax and Ideogram workflows — Constant-Tower9662 · 2026-10-02
- MIT's InstructMesh fixes AI-generated 3D models for real fabrication — nordicinst · 2026-10-02
- Suno launches Speech beta, the first audio model generating voice with matching music — suno · 2026-10-02
- MiniMax H3 continuation clips break context: prompt generator has no memory of prior scene — SnooMacaroons1365 · 2026-10-02
- MiniMax H3 RefMods deep dive: video, audio, motion & style with full ComfyUI workflows — Citadel_Employee · 2026-10-02