I love the concept behind HCB and the amount of effort everyone at Contra Labs is putting into testing and comparing models through pure taste and creative direction.
So I decided to give it a shot myself and create these two clips using the exact same prompt, but with different models.
Curious to know which one works better as a hook according to you?
5
21
3.5K
Built an immersive fantasy portfolio where every section becomes part of a living world instead of another webpage.
Designed the entire experience in @flowstep_ai using Flowstep + MCP (Antigravity).
Connected Flowstep MCP with Antigravity, then iterated directly on the canvas using promptsβrefining layouts, components, navigation, interactions, and the design system without rebuilding screens from scratch.
Created reusable design tokens, cinematic UI, glass navigation, parchment components, handcrafted cards, and a consistent visual language across every scene.
Once the prototype was finalized, I replicated the entire Flowstep canvas into a Next.js application and polished interactions with modern animation libraries before deploying.
The workflow:
πΏ Idea β Prompt Flowstep
π¨ Refine with MCP (Antigravity)
𧩠Build reusable design system
πΌ Iterate every screen on the canvas
β Replicate to Next.js
β² Deploy on Vercel
π‘ 9 connected fantasy scenes
β¨ 1 immersive portfolio
Project link : https://portfolio-6asf.vercel.app/
flowstep link : https://app.flowstep.ai/file?activeFileId=dd5c563d-33a1-4b37-8b24-3669efac9652
1
199
"We pay for ChatGPT and nobody uses it": a 2-week AI adoption sprint.
The problem: a company buys ChatGPT Team or Gemini licences, runs one demo, and three months later only two people log in. Staff don't know what to use it for in their actual job.
The fix: short interviews to find the 5 tasks each role repeats every week, copy-ready prompts for exactly those tasks, and two hands-on sessions where people use them on real work.
Includes: a use-case library by role, a one-page "what not to paste into AI" guide, and a 30-day usage check-in. (Sample build to show the approach.)
0
7
The Forecast is a 40-second psychological thriller short film built entirely on Renoise. The story follows a South Asian data analyst working late in a dark room, intensely focused on completing a critical forecast. Shot by shot, we see him type, analyze, and finally lean back with a satisfied smile as the screen flashes "FORECAST COMPLETE" β he believes he's done something meaningful. Then comes the twist: the camera pulls back to reveal he was never in an office. He's been sitting inside a glowing glass simulation pod the entire time, surrounded by scientists in white lab coats quietly taking notes on tablets β observing him like a lab subject. The final black screen delivers the gut punch: "Every forecast you trust... was made by someone who never knew they were being watched." Built using Renoise Canvas with Seedance 2.0 and FacePass for consistent character identity across all scenes.
Tool used: Renoise AI β renoise.ai
(http://renoise.ai)#RenoiseChallenge
3
1
312
π Scaling Your AI Models or Data Operations? Let Our Team of 40 Handled the Heavy Lifting!
Hey Contra Community! π
Are you struggling to keep up with massive datasets, messy annotations, or slow image verification pipelines?
If you are an AI startup, tech agency, or enterprise looking for a reliable partner to handle high-volume data tasks, we are here to help! I manage a structured, fully trained team of 40 data annotation and QA specialists ready to deploy immediately.
βοΈ What We Bring to the Table:
Massive Volume, Fast Delivery: With 40 people working in sync, we can process thousands of images, posts, or data rows daily without breaking a sweat.
Flawless Quality Control: We donβt just annotate; we implement a strict internal QA layer to ensure 99%+ accuracy before final delivery.
Versatile Expertise: Experienced in Image QA/Review (like matching AI outputs to inputs), Data Labeling (Bounding boxes, tagging), and Large-scale Data Entry (CSV/Sheets management).
πΌ Flexible Setup:
Whether you have a 1-week urgent sprint or need an ongoing monthly workforce, we scale according to your project requirements.
Let's remove the bottleneck from your operations. DM me directly to discuss your project requirements or to set up a quick sample/pilot task! Letβs build something great together. π
0
158
Fluffless: Deterministic AI Proposal Engine
Project Overview / TL;DR
A high-performance Next.js web application that forces LLMs to write hyper-direct, senior-level technical pitches by anchoring generation to a persistent, client-side profile vault.
The Problem:
The freelance market is flooded with generic, AI-generated proposals that read like bloated cover letters. Clients ignore them because they lack technical depth and focus on "marketing fluff" rather than execution. Additionally, existing tools create massive UX friction by forcing users to manually re-type their past work experience for every single pitch.
The Solution:
I architected and deployed Fluffless to solve both the output quality and the user friction. I bypassed standard AI wrappers to build a constraints-based engine that only generates technically sound, highly specific proposals.
Key Technical Execution:
Client-Side "Profile Vault": To eliminate repetitive data entry, I engineered a persistent LocalStorage architecture. Users input their raw technical resume once, and the application silently passes this base context to the AI for all future generations, reducing user input to just the target job description.
Heavyweight AI & Prompt Engineering: I integrated the Groq API utilizing the massive Llama-3.3-70B reasoning model. I engineered a strict, one-shot prompt architecture loaded with negative constraints to completely strip out default "AI-speak" (e.g., words like leverage, delve, seamless) and force a rigid 4-step psychological sales structure (Hook -> Proof -> Execution -> Technical CTA).
Native Streaming Architecture: When the standard Vercel AI SDK introduced silent failures due to protocol mismatches, I bypassed the abstraction entirely. I built a custom, native JavaScript fetch and TextDecoder implementation to guarantee 100% stable, low-latency text streaming directly to the UI.
My Role:
Lead Full-Stack Developer & Prompt Engineer
Tools & Technologies:
Next.js (App Router)
React
Tailwind CSS
Groq API (Llama-3.3-70B)
Vercel
Project Links:
Live Site: https://fluffless.vercel.app/
4
3
150
Built an approval-first AI workflow for LeadLens CareOS.
The system prepares customer-facing actions such as inactive-patient follow-ups, but nothing is sent automatically. Each action includes the exact payload, context, and expected impact, then waits for human approval before execution.
This is the type of AI automation I focus on: useful enough to reduce repetitive work, but controlled enough that businesses keep oversight of important actions.
Built with AI workflows, structured logic, approvals, CRM data, and business-process automation.
0
60
π¬ The Library of Rain
For the Renoise Challenge, I created a cinematic anime short inspired by the emotional storytelling of modern Japanese animated films. The story follows Ren as he discovers a mysterious abandoned library where every raindrop holds someone's memories. As the memories fade, he opens the final book and realizes the library has been writing his own story all along.
Created with Renoise, using consistent character references, cinematic keyframes, and AI-generated animation to achieve a feature-film anime aesthetic with immersive lighting, rain effects, and orchestral storytelling.
1
2
230