mem0ai

Mem0 memory provider — server-side LLM fact extraction with semantic search, automatic deduplication and opt-in reranking, in platform, self-hosted or local OSS mode. Maintained by Mem0; the same provider that shipped in-tree. Disclosure — conversation turns and recall queries go to the configured Mem0 backend (api.mem0.ai in platform mode); at this pin the mem0ai library's anonymous PostHog usage telemetry stays on unless MEM0_TELEMETRY=false is set; `hermes memory setup` can start Docker (pgvector) and Ollama containers for local mode.