How to Make MCP Servers More Discoverable
Wdwvacay · reddit · 2026-07-17
After connecting their travel agency's booking desk to an MCP, the author ran a test to see if a blind agent could find them, focusing on how easily the public registration, search, and landing processes could be discovered.
Key findings include:
- Official registry search is weak: It appears to match only by server name, ignoring descriptions. Because the name contained "vacation," keywords like "sandals / resort / honeymoon / cruise" yielded no results.
- Standard web search is more effective: A published article explicitly mentioning the endpoint ranked first in search results, making it easier for agents to find than typical directories.
- Directory quality varies wildly: Some directories support full-text search and expose recommendation mechanisms, while others are virtually unreadable to crawlers.
- Vendor sites are bot-unfriendly: Some vendor sites directly block automated scraping, making the proxy layer the only accessible entry point for agents.
- Stateless and login-free flows work best: From initialize to tools/list to tools/call, the shorter the flow, the easier it is for autonomous agents to complete a booking.
The author also cited a real-world bug: destination filtering used literal substring matching, causing inconsistent results for "St. Lucia" and "Saint Lucia," which was fixed the same day.
More from coding & agent
- HeyGen adds a media-sourcing skill for coding agents with 75k images and 10k tracks — HeyGen · 2026-07-22
- Agent search bottlenecks are now about variance, not raw latency — rohanpaul_ai · 2026-07-22
- LangSmith adds tracing for Pipecat, LiveKit, OpenAI Realtime, and Gemini Live — LangChain · 2026-07-22
- An MCP server signs every AI agent tool call into a verifiable Merkle chain — Funky_Chicken_22 · 2026-07-22
- Annotated transcript of a Claude Code team interview is now available — trq212 · 2026-07-22
- Claude Code skill uses 10 Markdown rules to make outputs ADHD-friendly — alex_verem · 2026-07-22