I build AI-native internal tools, workflow automation systems, desktop utilities, browser tooling, and MCP-style agent integrations for real operational work.

My strongest work sits where business process meets software: repeated manual tasks, fragile handoffs, reporting friction, support bottlenecks, and tools that teams actually need to use every day.

I turn messy operations into reliable tools — then use AI agents, automation, and custom software to make the work faster, clearer, and easier to repeat.

• AI tooling and agent workflows — MCP servers, local agent bridges, browser/tool surfaces, context systems, and Claude/Codex-oriented workflows.
• Workflow automation — n8n-style automations, reporting pipelines, ClickUp tooling, API integrations, and internal process systems.
• Desktop productivity apps — Windows/.NET/WPF tools for dictation, transcription, local-first utilities, and human-in-the-loop workflows.
• Browser and capture tools — Chrome extensions, DOM-to-Markdown exporters, page capture utilities, and web-to-AI handoff workflows.
• Business systems — practical software for support operations, documentation, reporting, task tracking, and team execution.

I'm focused on AI-assisted operational software: tools that help humans and agents work inside real systems instead of toy demos.

That includes:

• local machine-control gateways for AI agents
• MCP servers and agent-accessible tool surfaces
• AI dictation and transcription products
• ClickUp / workflow / reporting automation
• Chrome extension tooling for research and capture
• agent-first design and implementation workflows

Some of my strongest work is private or internal, but the build themes are consistent:

Local machine-control gateway for AI agents.
A Windows automation bridge for agent clients, combining MCP, UI Automation, screenshots, raw input fallback, verification logic, locator confidence, audit logs, and local safety controls.

AI voice dictation and transcription for Windows.
A WPF desktop app with global hotkeys, floating capture UI, multi-provider transcription, prompt profiles, local SQLite history, and workflow-specific transcription modes.

Agent-first design canvas.
A desktop design surface that gives coding agents a structured canvas, layout tools, visual verification loops, and JSX export paths instead of forcing them to design blindly in prose.

ClickUp CLI + MCP server for agent workflows.
Tooling for Claude Desktop, Claude web, and Claude Code workflows, with connector-specific behavior, compact tool descriptions, attachment-routing safeguards, and practical MCP hardening.

Chrome extensions, real-browser agent bridges, DOM-to-Markdown exporters, semantic memory systems, and utilities that move web context into AI workflows with less copy-paste friction.

flowchart LR
    A[Operational friction] --> B[Map the workflow]
    B --> C[Find repeated manual work]
    C --> D[Build internal tool or automation]
    D --> E[Connect APIs, data, and UI]
    E --> F[Add AI where it improves speed or judgment]
    F --> G[Make it usable in the real workflow]
    G --> H[Reduce friction, support load, and execution time]

    D --> I[Desktop apps]
    D --> J[Browser tools]
    D --> K[MCP servers]
    D --> L[Reporting pipelines]
    D --> M[Workflow systems]

AI tooling · workflow automation · internal tools · MCP servers · agent workflows · business systems · desktop apps · Windows automation · browser extensions · Chrome extensions · Python automation · TypeScript tooling · C#/.NET · WPF · n8n · ClickUp automation · OpenAI · Anthropic Claude · Gemini · OpenRouter · agentic systems · operational software

Automate the routine. Systemize the messy. Give humans and agents better tools to do the work.

• You need internal tools that remove repeated operational pain.
• You are building AI-assisted workflows and need practical tooling around them.
• You want browser, desktop, or MCP surfaces that make agents more useful in real work.
• You care more about shipped systems than buzzwords.