SplitBrain — Agent Card

🧠 SplitBrain: Two AIs think for you. You decide in ~10 minutes.

Price: $69 — one-time purchase, lifetime use

The problem you know too well

You start a brainstorm with your AI. At first, it's productive. Then it goes in circles. You rephrase, refine, add context. An hour later, you've burned tokens and your idea isn't any clearer than when you started.

The real cost isn't the API. It's the mental fatigue of being the only one pushing the thinking forward. But the API cost is real too — a 45-minute unstructured brainstorm with a top-tier model burns $2-5. Do that 10 times a month, you're lighting $500/year on fire for diminishing returns.

The solution: two brains are better than one

SplitBrain orchestrates a dialectical brainstorming sprint in 5 rounds max (~10 min typical) between two complementary AI models:

Result: in 3 phases and 5 rounds max, you get a structured file with all decisions, blind spots identified, and an actionable synthesis.

How it works (3 phases, ~10 min typical)

Phase 1 — Mapping
Your primary AI interviews you. 2-3 targeted questions. No fluff.

Phase 2 — Orthogonal Exploration
A second model analyzes the brainstorm and finds what logic didn't catch. Not a contrarian — an explorer that widens the field.

Phase 3 — Synthesis
Both perspectives are merged. You walk away with a clear plan and flagged open items.

Auto-stop if:

• You say "stop"
• A round brings nothing new
• 5 rounds are reached (hard cap)

What you actually get

✅ A structured Markdown file in your vault — key decisions, full Q&A log, explored angles, open flags
✅ The SplitBrain architecture: two AIs collaborating instead of monologuing
✅ A calibrated sprint: 80% of good ideas emerge in the first 2-3 rounds. This skill stops when signal fades, not when you're exhausted
✅ Token-optimized by design — SplitBrain pushes hard for 2-3 high-value rounds, then stops. No "one more question." No polishing forever. You take action instead of burning credits.
✅ Battle-tested: the creative brain has already corrected blind spots that logical analysis alone would never have caught

The "3-round rule" (why this works)

Brainstorming follows a logarithmic curve. Round 1 maps the territory (~50% of total value). Round 2 goes deep on critical branches (~80% cumulative). Round 3 catches blind spots (~95%). Rounds 4-5 exist as a safety margin — occasionally a complex topic genuinely needs the extra depth, but most sessions will stop at round 3 naturally.

SplitBrain is built on this curve. It pushes hard for 2-3 high-value rounds across two complementary models, then stops when signal fades. The 5-round cap is a ceiling, not a target — it prevents runaway sessions without cutting off genuine depth.

The math: a 45-minute unstructured brainstorm costs $2-5 in API fees. Do that 10 times a month, that's $500/year. SplitBrain gets you 95% of the value in ~10 minutes for pennies. The skill pays for itself in under two months of use.

Who is this for?

Content creators brainstorming video concepts, content series, or format ideas — validate your creative direction in a few minutes instead of chasing your own tail for an hour.

Solopreneurs & indie hackers who need to stress-test business ideas without a co-founder or committee. The creative brain acts as your second opinion.

AI power users who are tired of the monologue with their LLM. If you've ever felt like you're the only one pushing the thinking forward, SplitBrain fixes that.

Developers & architects designing systems, workflows, or technical architectures. Before you code, let two AIs find the blind spots in your design.

Writers & strategists outlining articles, video scripts, or marketing campaigns. The creative brain catches the angles your logical outline missed.

"I ran SplitBrain on an automated video editing workflow I was designing. My primary model helped me map out two separate pipelines — one for short-form, one for long-form content. Then the creative brain caught a blind spot: I was about to build two incompatible factories. It suggested a shared 'attention engine' core first, with each format as a different assembly recipe on top. That single insight saved me weeks of building the wrong architecture."
— Early adopter, content automation engineer

What's included

• SKILL.md — The main skill (complete orchestration logic)
• README.md — Installation, configuration, and how to adapt to your stack
• references/qwen-api-call.py — Clean Python helper for API calls
• Works with any model: DeepSeek, Qwen, Claude, GPT, local models via Ollama
• Works with any API: OpenRouter, Anthropic, OpenAI, OpenAI-compatible

Technical requirements

• Hermes Agent installed
• One OpenRouter API key (free, no commitment)
• 2 minutes to install

Why $69?

A Notion template costs $20-30. A prompt pack costs $30-50. SplitBrain is neither — it's a multi-model orchestration system that turns your AI agent into a thinking partner.

At $69, that's about 2-3 months of ChatGPT subscription. For a tool that improves every brainstorm you'll ever run. For life.

The skill pays for itself in under two months just on token savings alone.

Ready to give your AI a second brain?

Created by Capy Kamy