Finding your first 100 users with an agent-tool router

Problem

A founder's single most important job early on is finding and understanding their first users, and the tactics that work don't scale. Doing the research also means juggling many tool subscriptions and API keys, which is painful to wire into an agent.

Approach

Shengkun built Monid — an 'open router for agent tools' that gives an agent access to thousands of tools through one connection, with the agent paying per call (x402/MPP) and a $1 free credit good for roughly 500 requests. He then used his own product to find his first 100 users: via a skill/CLI/MCP in Claude Code, he had the agent mine competitor complaints on Reddit and Twitter (demoing on Lovable, filtered to the past 24 hours), draft outreach DMs, set a cron to monitor all competitors daily, and analyze competitors' growth strategies across platforms. The agent reads each tool's per-call pricing and picks a provider before running.

Results

Instead of generic LLM output, the agent surfaced real, timestamped complaints (some only hours old) and actionable growth insights — e.g. that non-English (Chinese/Japanese) influencer videos were driving growth for competitors, and that mid-sized creators (50k–200k) outperformed mega-influencers. Shengkun keeps a human in the loop, reviewing and sending DMs himself to avoid spam. He cited other users applying the same pattern: e-commerce product selection (Amazon data plus social signals), investors sourcing founders, and a 20k-subscriber YouTuber filtering which comments are worth answering.