HookLab AI

Demo

About

The problem it solves HookLab AI solves a painful creator bottleneck: most posts fail (or underperform) because the hook (the first line) doesn’t land — and creators waste tons of time iterating blindly. It also tackles a second friction point for Base/Farcaster creators: onchain + wallet friction makes experimentation feel risky, unclear, and slow. With HookLab AI, creators can go from: “I want to post” → “I have a strong post ready” in minutes, with a workflow that’s intentionally designed around what actually drives engagement. 🚀 🎯 What people can use it for 1) 🧠 Generate hooks built for engagement (not vibes) HookLab looks at what’s working on Farcaster (Base channel) and uses a measurable heuristic to evaluate hook performance: Hook Strength Formula hook_strength = (likes + recasts + replies) / follower_count This makes trend-based writing less random and more data-guided ✅ 2) 🙈 “Blind Hook Selection” to prevent wasting time Creators often over-edit the wrong thing (body, formatting, details) when the real issue is: the hook didn’t work. HookLab fixes this by enforcing focus: - Only shows hook previews first (first line, ~≤120 chars) - Full caption is hidden until a hook is chosen - Full caption is only revealed after a user commits to one This makes writing faster, safer, and more efficient ⚡ 3) 🧩 Intent-based caption generation (how creators actually think) Instead of “generate a caption” with no direction, HookLab starts from real posting intent, such as: 📢 Product update 🚀 Launch announcement 🧵 Thread opener 😂 Meme / culture post So you get outputs that are actually usable — not generic filler. 4) 🖼️🎬 Turn captions into media instantly (image & video) Once a caption is revealed, creators can expand into assets: - Generate image based on the caption context 🖼️ - Generate video (recommended: vertical 5–10s) 🎬 This turns “writing a post” into a full content kit (text + visuals) ✨ 5) 📌 Posting is streamlined (copy or deep-link) After unlocking: 📋 Copy the caption 🔗 Tap “Post to Base” to open composer prefilled No extra friction, just post faster ✅ 6) 💳 Credits model that feels fair + low risk HookLab reduces fear of paying for nothing: - New wallets start with 5 free credits 🎁 - Browsing hooks is free 👀 Credits are spent only when you unlock final value (caption reveal, image, video) So creators can explore freely, and only pay when something is worth using. ✅ 7) 🧬 Hybrid onchain/offchain that’s practical & defensible HookLab balances legitimacy + UX: ✅ Premium verification can be onchain ⚡ Quota/credits can be offchain for a gas-less experience This makes it hackathon-ready and user-friendly at the same time.

How it's made

Challenges we ran into 🧱 Challenges I ran into (and how I solved them) 1) 💥 Making credits fair (trust issue: “paid but didn’t receive”) This was the most important constraint: - Credits must be deducted only when a user unlocks value - If payment fails, credits must not be deducted - If unlock fails after payment, system should retry or restore credit ✅ How I solved it : - I separated preview state vs unlocked state - I treated unlock as the only event allowed to change credit balance - I designed retry / restore behavior to protect user trust 2) 🕒 Avoiding accidental charges during slow generation (especially video) Video rendering can take time, which creates risk: - Charging early = user feels scammed if result isn’t ready ✅ How I solved it : - I designed a clear lifecycle: - queued → rendering → ready → unlock - Credit is only spent when the asset is actually ready 3) 🔒 Enforcing “Blind Hook Selection” without leaking the full caption Blind selection is not just UI—it’s a rule: - Hook previews visible - Full caption locked - Reveal only after selection + unlock ✅ How I solved it : I structured output into two layers: - hook preview payload (free) - full caption payload (locked) - So the product forces the correct behavior: choose the hook first. 4) ⛓️ Hybrid onchain/offchain truth (premium vs quota) This is a classic Web3 tension: - Onchain is credible ✅ - Offchain is fast and practical ⚡ ✅ How I solved it : - Premium verification reads onchain state - Credits/quota stays offchain for UX - I avoided overclaiming “everything is onchain” (judge-safe) 5) 🧩 Ensuring “Post to Base” only uses unlocked content Deep-link posting is powerful but risky: - If it posts locked content, it breaks the credit model ✅ How I solved it - The “Post to Base” action is gated behind unlocked state - Composer payload always comes from the revealed caption, not previews 6) 🎥 Building a demoable flow under hackathon constraints Hackathon demos must be: - fast (under 60s) - instantly understandable - visually clean + no confusing branches ✅ How I solved it I designed a linear “game-like” flow: - 🔌 Connect wallet - 🎯 Choose intent - 👀 Browse hook previews (free) ✅ Select hook + unlock 🚀 Post to Base