File No. II · CheapestBuy.AI
Case Closed

The Case of the Vanishing Discount

RoleSolo Build
Timeframe2025
InstrumentsPython · CrewAI · Mistral AI · Solana Pay
$ EXHIBIT B — AGENT NEGOTIATION NETWORK

The Complaint

What Brought Me to the Case

Every online shopper knows the ritual: a dozen browser tabs, no unified way to compare listings across retail categories, and by the time you've found the best deal it may already be gone. The complaint was simple — nobody should have to do detective work just to avoid overpaying.

What was missing wasn't another comparison site. It was a verifiable system that could discover and optimize deals automatically, and then actually let you pay for what it found — without handing your card details to yet another platform.

The Investigation

How the Case Was Solved

I engineered a multi-agent shopping platform using CrewAI workflows to coordinate the investigation: agents that discover listings, verify them, and optimize across diverse retail categories, all driven by Mistral AI's LLMs for the reasoning in between.

For checkout, I integrated Coral MCP and Solana Pay, so the platform could complete secure, on-chain payments the moment the best deal is confirmed — merging conversational AI directly with Web3 payment infrastructure rather than bolting payments on afterward.

The result is a system built to be verifiable end to end: every agent's discovery and optimization step is auditable, and every checkout settles on-chain.

  • CrewAI multi-agent workflows for deal discovery and optimization
  • Mistral AI LLMs powering agent reasoning across retail categories
  • Coral MCP integration bridging conversational AI and payments
  • Solana Pay for secure, on-chain checkout

The Verdict

Closing the File

A verifiable multi-agent shopping platform that discovers and optimizes deals automatically, then settles payment on-chain — closing the loop from 'found a deal' to 'paid for it' in one system.