MCPqxbin

Pikk Product — MCP server for the QxBin Prompt Generator
Not open source. Source-available under commercial terms.

Probabilistic prompt engineering using Binary Probability Matrices.
Room-temperature classical cubit logic that evolves and collapses optimal prompts.

Owned and developed by Rupesh Malpani / pikk.company

Core framework: qxbin

---

Live Endpoint (Railway)

Public MCP endpoint:

https://web-production-f368a6.up.railway.app/mcp

Health check: https://web-production-f368a6.up.railway.app/health

Add to Grok

1. Open → https://grok.com/connectors
2. New Connector → Custom
3. Paste: https://web-production-f368a6.up.railway.app/mcp
4. Save

Then say:

Use the QxBin prompt generator on this goal: …

Tools

Tool	Mode	Description
generate_optimal_prompt	Edge	Single cubit, fast, seedable
generate_ensemble_prompt	Cloud	Parallel ensemble, picks global champion
generate_controlled_prompt	Auto / Edge / Cloud	Controller — auto-selects tier or force mode
refine_prompt	Hybrid	Automatic improvement — continues live matrix evolution with feedback + confidence
list_prompt_atoms	—	Inspect the atom library

---

Text Controller (simple)

Just type the goal.

python -m mcpqxbin.text_controller "Design a landing page for Pikkstops"

Design a landing page for Pikkstops
edge Write a product brief
cloud Compare three GTM strategies
Design X | Bootstrapped, under 180 words
atoms
help

File: src/mcpqxbin/text_controller.py

---

What it does

Treats prompt construction as a multi-dimensional probability landscape:

1. Atom library — Role / Reasoning / Structure / Quality / Constraint
2. Binary Probability Matrix — spatial grid of fractional probabilities
3. Superposition — bias**n + (1-bias)**m steers probability mass
4. Measurement — collapses to the highest-leverage prompt

QxBinController (generate_controlled_prompt) adds auto tier selection:

• short / simple goals → Edge
• long / multi-view goals → Cloud ensemble

Automatic improvement (refine_prompt): after a first generation, feed the same goal + feedback + confidence. The Binary Probability Matrix keeps evolving (Hybrid-Loop style) instead of cold-starting. Iterate until the score / lean is strong enough.

---

Local Use (Testing / Non-Commercial Only)

uvx --from git+https://github.com/pikk-qxbin/MCPqxbin.git mcpqxbin

or

pip install git+https://github.com/pikk-qxbin/MCPqxbin.git
mcpqxbin

Claude Desktop / Cursor config (stdio):

{
  "mcpServers": {
    "mcpqxbin": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/pikk-qxbin/MCPqxbin.git", "mcpqxbin"]
    }
  }
}

---

Example

Tool call: generate_controlled_prompt

goal: Write a sharp cold email to Mark Cuban about Pikkstops + QxBin
extra_context: Bootstrapped, under 180 words, first principles only
mode: auto
bias: 0.84

Returns a fully structured, high-signal prompt ready for any LLM.

Then improve it:

refine_prompt
goal: same
feedback: stronger CTA, more first-principles, under 150 words
confidence: 0.86

---

License & Commercial Terms

This is a Pikk for-profit product.

Source code is made available so others can inspect, learn from, and experiment with it.

• Free for personal testing, education, research, and non-commercial internal use.
• Commercial use (any product, service, API, SaaS, or monetized deployment that uses or derives from this code or the underlying QxBin logic) requires a commercial license or revenue-share agreement with pikk.company.
• Standard terms include a 51% revenue share on gross revenue from qualifying commercial products. Enterprise and strategic deals are negotiable.

See the full LICENSE for exact terms.

Contact: @rupeshmalpani or [email protected]

---

pikk.company — Building the probabilistic compute layer for the real world.