Minji Kim

Building MCP tools and data APIs for autonomous AI agents — capabilities that Claude, ChatGPT, and other agents call directly, metered per use.

AI product & strategy builder based in Seoul 🇰🇷, working bilingually (EN/KO). I ship agent-callable tools across Apify, Cloudflare Workers, and Containers, and turn frontier LLM capability into product.

• MCP tools for AI agents — exposed over Streamable HTTP, invoked by agents rather than humans through a UI.
• Data & automation Actors — scraping, browser automation, and Docker workloads too heavy for an edge runtime.
• Usage-priced endpoints — pay-per-event / pay-per-result, so an agent can evaluate cost before it invokes. No flat rental lock-in.

Area	What it does
Agent tools (MCP)	Callable capabilities for LLM agents — structured lookups, domain engines, calculators
Data extraction	Reliable structured scraping with clear per-result pricing
Compute services	Docker-based processing that doesn't fit serverless edge

• Agent-first I/O — tight schemas, predictable output, low latency, small context footprint.
• Transparent pricing — priced per event/result; cost is knowable before a run.
• Maintained — monitored, versioned, and kept current.