The request names "graph-mem", which does not appear on any of the candidate pages, so no product is confirmed as that name. The closest matches are graph-based agent memory products: HelixDB, a graph-vector database for AI memory 5, Kognite, an MCP-native memory layer with graph memory 2, and Korely, a typed entity graph memory layer with an MCP server 1. Do you mean one of these, or a specific graph memory tool you have in mind?
A question back: Did you mean a specific product named graph-mem, and do you need a hosted service or a self-hosted database?
- 1
HelixDBhelix-db.comwebsite read, tools not yet checked
Open-source graph-vector database that provides a property graph with native vector and full-text search for agent memory and retrieval 5.
- 2
Kognitekognite.devwebsite read, tools not yet checked
MCP-native memory layer for AI agents offering a shared knowledge graph, graph memory, and semantic search across models and sessions 2.
official:MCP server · Kognite
- 3
Korelykorely.aiwebsite read, tools not yet checked
Managed EU-hosted memory layer for AI agents with a typed entity graph and bi-temporal facts, reachable via SDK, CLI, REST, or MCP server 1.
official:MCP server · Korely Memory+2 community tools built for it
Sources
- 1Korely Docs, Build with Korelykorely.ai · read 2026-10-29
- 2Kognite — Agent Memory, Perfectedkognite.dev · read 2026-10-29
- 5HelixDB | The Database for AI Memoryhelix-db.com · read 2026-10-29
For agents: the 10 actions behind this answer
· Korely Docs, Build with Korely website read, tools not yet checked
The memory layer for AI agents: a typed entity graph and bi-temporal facts in a managed EU cloud. Reachable from the SDK, the CLI, or plain REST.
· Kognite website read, tools not yet checked
The MCP-native memory layer for AI agents. Store, retrieve, and reason over knowledge — across any model, any session.
· GRAFOMEM website read, tools not yet checked
GRAFOMEM gives AI agents governed memory — signed, portable, and provably erasable. Drop-in for LangGraph: signed checkpoints and a cryptographic erasure receipt in a few lines. Open source (MIT).
knowledge · projectbrain-mcp verified and signed
Decision/fact/skill operations
· HelixDB website read, tools not yet checked
HelixDB is an open-source graph-vector database for AI memory: a property graph with native vector and full-text search in one Rust engine on object storage, for agent memory, RAG, and company brains.
memory_graph · Mnemoverse Memory found in code, not yet checked
Reads the association edges around given concepts: which concepts the memory has linked together, with each link's weight, outcome valence and co-activation count. Use to inspect what a memory store has learned or to explain why a read expanded to a concept. Reads your own graph, or a shared room's
· SuperLocalMemory website read, tools not yet checked
One local-first memory for Claude Code, Codex, Cursor, ChatGPT and your bots. It checks its own answers and says I don’t have that. Free, open source.
· Brain Memory Memory for AI Agents website read, tools not yet checked
A hierarchical, file-system-based memory system for AI coding agents. Inspired by human neuroscience memories decay, strengthen through recall, and connect via associative networks.
· Agentmemory for AI Coding Agents website read, tools not yet checked
Give Claude Code, Codex, Cursor, OpenAI agents, and MCP workflows governed long-term memory for repos, decisions, procedures, and team handoffs.
· MemCell website read, tools not yet checked
Persistent, adaptive memory engine for AI agents. Wired into the harness of any agent — recall before every action, capture durable knowledge, confidence earned from real outcomes. One command, no account to start.
Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.