
upload | 跨 AI 的长期记忆
A shared memory layer that persists your decisions, preferences, and projects across different AI tools and devices.
upload — 跨 AI 的长期记忆层,让 ChatGPT、Claude、Codex、Hermes、OpenClaw、WorkBuddy 等共用同一份记忆,换工具换设备都接着上次继续,注册即用免装插件
The full gallery
Tech stack
28 projects

A shared memory layer that persists your decisions, preferences, and projects across different AI tools and devices.
upload — 跨 AI 的长期记忆层,让 ChatGPT、Claude、Codex、Hermes、OpenClaw、WorkBuddy 等共用同一份记忆,换工具换设备都接着上次继续,注册即用免装插件

Create isolated, stateful development environments for coding agents and engineering teams.
@simskiiee · X

Upload and share files up to 1GB free with client-side AES-256 encryption, no signup required.
@ecom_jona · X
CARTOON ADS are taking over Meta right now!!!🔝 Made this entire ad with one workflow I built with Claude. No agency, no filming, no editing. These formats are converting higher than almost anything else right now. Download my workflow here:

A cloud Linux workstation with 60+ AI tools including Claude Code, Codex, and Gemini.
@CoderLuii · X
A machine for your coding agents.

Browser-based SSH client with server management and SFTP file access
CloudSSH — 打开浏览器即可连接和管理自己的所有服务器,连接信息云端同步,内置 SFTP 文件管理与 AI 运维助手 - [更多介绍](https://github.com/newbietan/CloudSSH)

Transfer files or text between devices using a one-time pairing code with end-to-end encryption and automatic cleanup.
@FrankFika · X
Still using a chat app to send files to yourself? I built USend: no sign-in, no app, no permanent share link. An 8-digit code moves a file or text between devices. Browser-side encryption. Auto-delete after receipt or expiry. Try it: #VibeCoding #OpenSource

Hosted JSON database for storing agent memory with REST and MCP connectivity
@StuSim · X
hey Adam, I run , lightweight agent memory

Sync data across devices in real-time with automatic security inspection via an API.
@mygridsyncdev · X
Just launched GridSync ⚡️ Sub-150ms zero-trust payload inspection & real-time sync for modern APIs. Try the free tier on RapidAPI or check out the docs: #buildinpublic #API #FastAPI

Write once, export to notebooks, slides, PDFs, PPTX, DOCX, or canvas.
@mr_wickedhacks · X
Write the doc once → get a notebook, slide deck, canvas, or resume from the same source. No rewriting for every format.

Share encrypted files with self-destructing links without signing up.
@cortexdrop · X
CortexDrop- first consumer file sharing built on IPFS. Drop a file, get a private link.

Plan, produce, review, and publish creator content all in one workspace.
@DaiwikShashi · X
Building Scribur — the operating system for in-house UGC teams. We help in-house teams manage campaigns, content ideas, scripts, submissions, revisions, approvals, creator communication, and performance tracking all in one place.

Annotate and redact files in your browser, then share with auto-expiring links.
TANTFENG · V2EX
周末到啦 介绍一个更加通用的 AI 导盲犬 Ampcode 推出的 Orbs 非常好用 但是反反复复提交图片 debug 会有点心累 现在远程 vibe coding 开发的场景越来越多了 agent 的识图能力也越来越强 但是如何给 agent 传递且描述准确的信息成了瓶颈 像 cursor design mode 这种好用的工具并不在大多数地方提供 尤其是画面动态变化的时候 或者步骤较多的时候可能需要多次标注以及关键帧引导才行 Markup Relay 作为 SUB&SUB 工作室第一个构想的开源产品提供了一个还在不断完善的思路 来作为 agent 的导盲和接力助手 让标注的内容能够被 agent 读懂 即便是视频 也可以方便的提取关键帧来描述画面的变动 https://tools.subnsub.com/relay 使用方法很简单 在桌面端浏览器内 你可以通过拖动支持的文件到窗口内进行快速编辑 如果是框选工具 在处理完毕后会对未框选的区域在浏览器端自动打码并上传至 cloudflare 的 R2 桶 并在设定时间内自动删除 https://i.