The problem: too many clients, too few discovery hooks We expose Supabase Edge Functions as MCP (Model Context Protocol) servers. The clients that hit them are heterogeneous — Claude Desktop, Codex CLI, Cursor, VS Code Continue, a couple of in-house browser extensions. None of them ship with a hard-coded "use WorkOS AuthKit, scope is tool:ai_chat, audience must contain urn:jibun:tool:<tool>" rec
When you have 5 unrelated questions, should you pack them into one message to the LLM, or send 5 requests simultaneously? Which is faster? Splitting into multiple independent parallel requests is almost always faster. This isn't a gut feeling — it's determined by the underlying inference mechanism of LLMs. Let's walk through the reasoning from first principles. To understand this problem, you firs