Supermemory Hits 19.1K Stars — Memory API Processing 100B Tokens/Month at Sub-300ms Latency

GitHub·medium signal

Supermemory, a serverless memory engine and API for AI applications, has reached 19,138 stars (+809 today). The platform processes over 100 billion tokens monthly with every query responding in under 300 milliseconds, serving as context infrastructure for AI agents with user profiles, memory graphs, and semantic retrieval. Following a March 6 service degradation incident (their first major downtime), the team has shipped significant infrastructure hardening.