reddit · SideProject · 11-08 22:50 63 分 ← 返回列表

I got tired of making OG images and favicons in Figma, so I built a tool that lets my AI do it

I build a lot of side projects, and every one needs the same marketing images: OG images, favicons, logos, blog post banners. Every time it's the same thing: look for inspiration, open Figma or Canva, move things around, export, repeat. So I built Render Pixels and now have Claude do all of that for me. How it works: you ask your AI for the image, it creates the design as HTML/Tailwind, and Render Pixels renders it into a pixel-exact image. It's an MCP server, so it works with any AI that

五维评分

付费意愿

需求真实度

个人可落地

竞争空间

新鲜度

AI 分析

痛点：独立开发者每做一个新项目，都要重复做 OG 图、favicon、logo、博客横幅，切换 Figma 或 Canva 反复调整、导出，耗时且枯燥。

目标用户：频繁发布 side project、博客和产品落地页的独立开发者、开源作者和小型内容团队。

证据：every one needs the same marketing images: OG images, favicons, logos, blog post banners. Every time it's the same thing: look for inspiration, open Figma or Canva, move things around, export, repeat. ... You get 50 free renders per month, then pay as you go.

变现：每月免费额度加按次计费；可升级为团队订阅，提供模板库、批量渲染和 CI/CD 集成。

MVP：用 Satori 或 Puppeteer 把 HTML/Tailwind 渲染成 PNG，提供一个 MCP server 和几个预设模板（OG 图、favicon 多尺寸、GitHub social preview），用配额计费快速上线。

竞品：Vercel OG / Satori、Bannerbear、Placid、htmlcsstoimage、ScreenshotOne、Canva

摘要：作者因为反复做 OG 图、favicon 和博客横幅太繁琐，做了 Render Pixels：让 AI 生成 HTML/Tailwind，再由 MCP server 渲染成像素级精确的图片。每月免费 50 次渲染，之后按次付费。痛点真实且普遍，但同类 OG 图和图片渲染工具已经很多，差异化只能落在“AI 原生 + MCP 集成”上。