📡 Idea Radar 每天 8:00 / 18:00 推送 · 18:00 生成精选
reddit · SideProject · 56 分 ← 返回列表

I got tired of re-uploading the same references for AI images, so I built this

Hey everyone! I'm a product designer, and I've been working on something that started as a workaround for my own AI image workflow. Every time I started a new image chat for the same brand, I ended up uploading the same product photos, explaining the same person, and describing the same setting all over again. I tried keeping reference sheets, but eventually I was spending more time managing them than actually making images. So I built Scenri, an open-source photo workspac

五维评分

付费意愿
45
需求真实度
50
个人可落地
82
竞争空间
42
新鲜度
70

AI 分析

痛点:做品牌/系列 AI 图像时,每次开新对话都要重复上传同样的产品图、人物和场景参考,管理参考资料的时间甚至超过真正做图的时间。

目标用户:需要长期保持视觉一致性的产品设计师、品牌设计团队、电商/广告内容创作者,以及高频使用 AI 图像生成的人。

证据:Every time I started a new image chat for the same brand, I ended up uploading the same product photos, explaining the same person, and describing the same setting all over again. I tried keeping reference sheets, but eventually I was spending more time managing them than actually making images.

变现:开源本地版免费+付费云同步/团队协作版;或按席位订阅的参考资源库托管服务,支持品牌资产库、历史版本管理与多人共享;也可作为插件对接 Midjourney/Flux 等付费额度。

MVP:做一个本地优先的参考资产管理工具:一键保存产品/人物/场景为可复用卡片,自动注入到主流 AI 图像工具的提示与上传流程中,附带历史记录和一键切换光照等参数,2 周内用 Node.js+Electron 或浏览器插件即可完成。

竞品:ComfyUI + 工作流模板、LoRA 训练/一致性工具、Midjourney Character/Omni Reference、Adobe Firefly 品牌资产库、各类 prompt/asset 管理工具

摘要:这是一位产品设计师因反复上传品牌参考图而自建的 Scenri,一个运行在本地的开源 AI 图像参考工作台,可保存产品、人物、场景并复用,还能通过 Codex CLI 复用已付费的 ChatGPT 订阅而无需额外购买出图额度。痛点真实且被个人验证,作者本人刻意规避额外付费说明价格敏感度高,付费意愿证据较弱,但做团队版/云同步的商业化空间存在。竞品(LoRA、ComfyUI、各家图像工具的原生参考功能)较多,需在本地化+免额度这一细分点切入。