Asad Jehan Zeb
Web developer and vibe coder
Ready for work
Asad Jehan is ready for their next project!
AgentDesk — The Control Room for AI Agents
AgentDesk is a control room for AI agents, built to be remixed as a real SaaS starter — not a single-screen chatbot demo.
Every run goes through one clean interface, AgentRuntime, with two sides: a fully deterministic demo runtime that needs zero API keys and works instantly, and a provider runtime that makes a real request to whichever model provider you've configured. Three providers ship today — OpenAI, Anthropic, and OpenRouter — each behind the same generic ProviderAdapter interface, so adding a fourth is one small file, not a rewrite. Credentials are read from server environment variables only; they never reach the browser, get stored in localStorage, or appear in logs, even when a request fails.
The Integrations and Settings pages don't fake this — they show live, server-verified Connected/Not-configured status per provider, so what you see is exactly what's true, not a static mockup pretending to be "connected."
Ten real screens: Overview, Agents, Run detail with a full execution trace, Tasks, Knowledge, Integrations, Analytics, Settings, in-app Docs, and a landing page — all in a distinctive warm, editorial design system (cream backgrounds, teal and terracotta accents) instead of the generic dark AI-dashboard look most agent tools default to.
Fork it, swap in your own domain's demo data, and you have a working foundation for an AI-agent product — not just a UI shell.
Live app
https://agent-desk--zuhaibbusiness1.replit.app
Repo
https://github.com/zanni098/Agent-Desk
0
74
AgentDesk Development Project
0
0
Gravity Flip drops you into a neon perspective tunnel where your own face is the character, and your mouth is the only control — open it to flip gravity and dodge your way to a higher score.
I built it entirely in Easy Lens with no code, combining the Game Creation and Face ML tools: I described the endless-runner loop and the neon synthwave world, wired the face avatar into the tunnel, and mapped mouth-open as the flip input. The rest was iteration — tightening the calibration screen, the scoring, and the difficulty ramp so runs stay short, punchy, and replayable.
Tools used: Easy Lens (Game Creation + Face ML Generator)
Tag: #easylenscontra
4
192
Bumper — the motion graphics tool I wanted to exist. My
AI video tools charge you per second because they run on GPUs. But a lower third isn't a generation problem, it's a rendering problem — 105 deterministic frames a browser can draw for free. Same input, same 105 frames, every time. So I built the thing that treats it that way: pick a preset, tune it on a real timeline with named layers and visible easing, drop your own footage underneath, then export a transparent overlay for your edit or the finished shot with the footage baked in. No credits. No subscription. No GPU bill to pass on to you.
Built in Flowstep and driven through the Flowstep MCP server from Claude Code — every screen prompted straight onto the canvas instead of pushed around by hand. 8 screens on one design system: New graphic → Add footage → Editor → Brand kit → Export, plus the landing page, sign up, and an alpha variant of the editor.
Flowstep file (public): https://app.flowstep.ai/file?activeFileId=22ceed96-9211-4b25-9fb2-81c452733386
Live site: https://bumper-site.vercel.app
Code: https://github.com/zanni098/bumper-site
X post: https://x.com/Elephanto_101/status/2083399037240381837
Attached: the full walkthrough, plus a recording of the Flowstep canvas behind it.
Finished this after the buzzer, so it's out of the running. Posting it anyway — I'd rather it exist here than nowhere.
0
160