🚀 Devlog: 23 Days Into Soft Launch

We're now 23 days into the soft launch of ManaTap.ai, and the project has already grown faster than we expected. What began as an early-access experiment is turning into a genuinely smarter, sharper deck-building assistant — thanks entirely to the players who've been testing, breaking, and refining the tool with us.

🔧 What's Changed Since Launch

A soft launch is where all the real learning happens. Here's what the last few weeks have brought:

🪲 Dozens of Bugs Fixed from Live Users

Everything from subtle UI quirks to deeper logic issues inside the analyzer has been ironed out.

If someone has reported it, chances are it's fixed or already in the patch queue.

Real users hit corners of the rules, formats, and card interactions no dataset ever catches — and that feedback loop has been gold.

🤖 The AI Has Levelled Up — Massively

The biggest leap forward this month has been in the intelligence of the deck analysis engine itself.

Across 23 days, we've improved:

• Archetype recognition — the AI is far better at identifying your deck's true plan.
• Synergy awareness — token, aristocrats, landfall, lifegain, enchantress, spell-slinger… each now gets more precise, on-plan suggestions.
• Format accuracy — fewer off-color, off-format, or banned recommendations.
• Ramp categorization — land ramp vs mana rocks vs dorks is now clean and consistent.
• Budget-friendly reasoning — clearer explanations about cost, upgrades, and affordable swaps.
• Teaching mode — responses have structure, clarity, and real, actionable examples.

Behind the scenes, hundreds of micro-updates to the system prompt mean the AI behaves far more like a real MTG player — not a card database.

The difference between Day 1 AI and Day 23 AI is night and day.

And we're not slowing down.

⚡ Faster Combo & Synergy Detection

Combo detection and synergy scanning now run significantly faster, with fewer false positives and better explanations of why something works.

This is still evolving, and you'll see more improvements in December.

💬 Better Reasoning Behind Card Suggestions

One of the biggest requests from early testers was:

"Explain your thinking."

So we've done exactly that.

Card suggestions now come with much clearer logic:

• Why a card fits your plan
• Why it pushes curve or budget too far
• Why something might be slow in your format
• What alternatives exist

The assistant doesn't just suggest cards anymore — it teaches.

💙 A Huge Thank You to Everyone Testing

Every deck you upload teaches the model something.

Every bug you report removes friction.

Every message shapes how ManaTap.ai evolves.

You're not just early users — you're co-builders.

Your fingerprints are already all over this project.

🧭 What's Coming Next

Based on feedback from the last 23 days, here's what we're focusing on next:

• Even deeper archetype tests (recursion, blink, voltron, treasure engines)
• Stronger format detection and legality checks
• More accurate land-base advice
• Tighter synergy detection and curve analysis
• UI polish and better onboarding
• Continued AI refinement through daily test evaluations

December is shaping up to be a huge month.

Ready to See How Much Smarter It's Become?

Try a deck you tested on Day 1 — the difference will be obvious.

🔗 Launch the Deck Assistant →

Feedback buttons are built into every analysis. Use them. We read everything.

Thanks again for being here during the messy, magic-making early stage of ManaTap.ai. The adventure's just getting started.