Test headlines with real human feedback.

You start with 30 credits.

Rate 1 set, earn 1 credit.
Submit a single headline, costs 3 credits.
Submit a batch (2-10 headlines), costs 2 credits per headline.
Keep a 3+ day streak, earn +0.5 bonus per rating.

That's it. Rate headlines from other people to earn credits. Spend them to test your own.

The mutual aid model. You can't just buy your way in. You earn credits by rating, which gives everyone else the data they need. No $200/month subscriptions. No free trials that convert to paid. Just trade attention for attention.

If you run out of credits, rate more sets. Your daily and hourly limits increase as you level up (Training-Stick starts at 20/hour, 50/day. Ta-Seti gets 100/hour, 500/day).

Why it's structured this way: Makes you actually look at other people's headlines instead of just submitting yours and leaving. The platform only works if people rate thoughtfully. Credits force skin in the game.

Your level determines how many headlines you can rate per hour and per day.

Level 1 (Training-Stick): 20/hour, 50/day.
Level 10 (Ta-Seti): 100/hour, 500/day.

Levels are based on two things: accuracy score and sets rated.

Accuracy measures whether you're actually reading the headlines or just clicking randomly. Control headlines test this. Obviously good content you should click, obviously bad content you should ignore. Click good stuff: +2 points. Click garbage: -5 points.

Your accuracy starts at 50%. Get to 60% and rate 50 sets, you hit Level 2. Keep going, eventually Level 10.

Higher levels = more rating capacity. It only makes sense to let people who pay attention rate more headlines. Random clickers stay at low limits. People who read carefully get access to more testing volume.

Each level has a weapon name because the attention economy is a battlefield and generic gamification is boring.

Level 10 requires manual approval from me. You'll need 85% accuracy and 2000 sets rated. At that point you can leave comments on headlines.

Your headline goes live immediately. No approval queue.

It gets shown to other users until it hits 100 impressions. Usually takes a week or two depending on how many people are rating.

Once it hits 100 impressions, marked "complete", you can see the click-through rate.

After 40 days, headlines get archived automatically. After 90 days, deleted. This isn't negotiable. The platform isn't a permanent storage system for every headline you've ever tested.

You can't see your own headlines in the rating pool. No gaming the system by clicking your own stuff.

There are 30 categories: technology, business, sports, gaming, etc.

When you submit a headline, tag it with 1-5 categories. "Tesla Cuts Prices" could be automotive + business + technology.

When you rate headlines, you see ones in your selected interests first. If there aren't enough in your interests, you get shown everything.

It only makes sense to ask for your interests so you don't skip over headlines where you weren't the target audience anyway. A tech person judging beauty content is useless data for everyone.

Every time you rate, you see 5 headlines at once. Pick the ones you'd actually click.

You can click 0-5. Doesn't matter. Just be honest about which ones would get your attention.

80% of the time, one of those 5 is a control headline testing if you're paying attention. You won't know which one.

Once your headline hits 100 impressions, you get the click-through rate.

That's it. No complex analytics dashboard. Just: X people saw it, Y people clicked it.

If your CTR is low, your headline didn't work. If it's high, it did. Use that to test variations before spending real money on ads.

This isn't a headline generator. That's the problem I'm trying to solve. AI advice stops working when everyone uses it.

Think of it as a "Decentralized assist for assist". The mutual aid model stays. You can't just buy your way past rating other people's work.

I'll introduce other features later, maybe sooner than you think. Just need to fullproof this one function. Secure it. And see that I'm solving a problem for many creators and not just me.