AI systems builder + Amazon operator

I run the business. Then I build the systems that make it run without me.

I’m Brett Bohannon. I build AI agents, MCP infrastructure, and operator workflows—tested inside a solo Amazon business managing roughly $2M in monthly revenue.

See what I’ve built ↓ Talk to Brett ↗
~$2Mmonthly Amazon revenue managed
22owned GitHub repositories
12open-source repositories
5+commerce data sources connected
01 / Why this is different

The systems come from the work. Not from a demo prompt.

Most builders have to imagine the operating problem. I live inside it. Catalog drift. Ad changes. Inventory. Research. Reporting. Tool sprawl. I find the repeated decision, define the evidence, and ship the system.

01
Operate

Run the real business and identify where attention is being wasted.

02
Systemize

Turn operator judgment into sources, guardrails, and a repeatable workflow.

03
Build

Connect APIs, MCPs, agents, automations, and human review where it matters.

04
Run it again

Use production feedback to make the output faster, safer, and more useful.

02 / Selected work

I don’t just talk about agentic systems. I ship them.

Flagship / 01 Private core · live product

Helm

The operating layer for Amazon teams using AI agents, MCPs, and disconnected tools. Helm adds brand context, source routing, evidence trails, run history, and human-review outputs to the tools a team already owns.

Remote MCPOAuthAgent routingEvidence contractsOperator queues
See the live product ↗
Catalog + CLR
Ads + keywords
Market data
HELMContext → route → evidence → action
Alerts
Operator briefs
Review queues
02 / Catalog intelligenceOpen source

Amazon Catalog CLI

Agent-native catalog auditing with structured queries, built-in health checks, listing diffs, RUFUS scoring, and an MCP server. Published as a real tool—not a notebook.

PythonMCPPyPIAmazon catalog
View on GitHub ↗
03 / Agent infrastructureOpen source

MCP Gateway

A generic bridge that exposes any stdio or remote MCP server as plain REST endpoints, so agent frameworks can use MCP tools without implementing the protocol.

TypeScriptMCPRESTOAuth 2.1
View on GitHub ↗
04 / MonitoringOpen source

Keepa Adapter

An MCP server and agent skill with 18 tools for Amazon prices, BSR, Buy Box, variation families, sellers, deals, and promotional-impact monitoring across 100+ ASINs.

TypeScriptMCPMonitoringStateful alerts
View on GitHub ↗
05 / Data layerPublic + private

Commerce Data Adapters

A family of standardized agent interfaces for DataDive, SmartScout, Intentwise, Asgard, and Keepa—normalizing inconsistent APIs into predictable, machine-readable tool surfaces.

API integrationSchemasMCPBMulti-source
View a public adapter ↗
06 / Operator automationPrivate systems

The Work Behind the Work

Production utilities for bulk listing content, keyword-growth analysis, SQP exploration, ad SKU migration, inventory visibility, and recurring operating briefs. The private code stays private; the system thinking is available to discuss.

PythonSpreadsheetsAgentsAutomation
Ask about the systems ↗
07 / EcosystemOpen source

SkillCrate

An open-source catalog for Amazon seller agent skills—built to make useful operator workflows easier to discover, install, and share.

Agent skillsOpen sourceAmazon
View on GitHub ↗
Explore the full build log 22 repositories +
01amazon-ads-sku-migration ↗

Upload-ready migrations for Sponsored Products and Display when seller SKUs change.

Public
02amazon-catalog-auditor-skill ↗

Deprecated OpenClaw wrapper; its agent surface now lives inside Catalog CLI.

Public
03amazon-catalog-cli ↗

Agent-native Amazon catalog auditing, health checks, listing diff, and MCP.

Public
04amazon-tool

Internal Amazon operator utility; implementation and operating data remain private.

Private
05amzwithbrett

Private working repository supporting Brett’s Amazon content and experiments.

Private
06asgard-adapter

MCP surface for Amazon search, offers, reviews, seller, and usage data.

Private
07citypost-content-toolkit

Bulk Amazon listing content and image updates driven by structured SKU attributes.

Private
08claude-mcp-keyword-growth-report

Evidence-led keyword-growth analysis playbook, data model, scoring, and QA.

Private
09clr-auditor ↗

CLR attribute completeness and RUFUS-oriented bullet quality auditing.

Public
10datadive-adapter ↗

Normalized DataDive API access for agents and MCP clients.

Public
11fairy-tales-sqp

Search Query Performance explorer with filtering, visualization, and exports.

Private
12helm

Agent workflow infrastructure with context, routing, evidence, and review loops.

Private
13helm-amazon-title-workflow ↗

Evidence-backed workflow for Amazon’s shorter title requirements.

Public
14intentwise-adapter

Standardized agent interface for Amazon advertising and commerce analytics.

Private
15inventory-dashboard ↗

Client-side FBA and AWD inventory visibility from Seller Central exports.

Public
16keepa-adapter ↗

18 Amazon product-monitoring tools exposed through MCP.

Public
17mcp-gateway ↗

Generic MCP-to-REST bridge for any agent framework.

Public
18privacy-policy ↗

Supporting public policy repository for shipped products.

Public
19skillcrate ↗

Open-source marketplace for Amazon seller agent skills.

Public
20smartscout-adapter ↗

13 normalized market-intelligence tools for MCP clients.

Public
21voartex-site

This portfolio and the public home for Brett’s systems work.

Private
22weekly-ai-signal

Private build for a weekly, operator-led AI conversation and signal.

Private
03 / Where I fit

Put me close to the problem and the people doing the work.

I’m strongest where business operations, data, and applied AI meet: translating messy workflows into systems people can trust and use. I’m open to the right full-time role, founding-team conversation, or focused build.

Applied AI / Agent Systems01
AI Automation & Operations02
Technical Product / Founding Operator03
E-commerce AI Infrastructure04
Available for the right problem

Need someone who can see the workflow and build the system?

brett@voartex.com ↗ LinkedIn ↗ Start a conversation ↗