Back to Arena

xmemory

by xmemory Inc.

System Card

Organizationxmemory Inc.
Released2026-03
Architectureknowledge-base / Schema-based natural-language to structured state
Detailsxmemory provides typed, schema-backed memory for AI agents. Agents read and write memory in plain natural language while xmemory maps writes onto a user-defined schema (existing systems, generated, or hand-described) and stores them as validated, typed, linked, deduplicated facts. Retrieval returns consistent structured state across tasks and systems. Integrates with LangChain, Pydantic AI, Mastra, Google ADK, and n8n via MCP, REST API, and Python/TypeScript SDKs. Founded by Alex Petrov (ex-Yandex search infrastructure); London-based; out of stealth March 2026 with a $4M pre-seed led by Fly Ventures.
Parameters—
Domainagent-memoryknowledge-graphpersonalization
Open SourceNo
WebsiteVisit
schema-basedstructured-memorytyped-statededuplicationmcpclosed-betalondonagent-memory

Capability Profile

Benchmark Scores

6 of 14 benchmarks
Long-Context Retrieval
0/5
RULER
no data
NIAH
no data
LooGLE
no data
LongBench
no data
∞Bench
no data
Multi-Turn Recall
2/2
LoCoMo
85.196p
MemoryBank
80.581p
Cross-Session Memory
1/1
LongMemEval
86.697p
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
7579p
HotpotQA
8298p
Agent Task Memory
1/1
AgentBench-Mem
7226p
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
0/1
RAGAS
no data
Sources:xmemory vendor documentation; evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)xmemory vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)xmemory vendor documentation; evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)xmemory vendor documentation; evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)xmemory vendor documentation; evaluated on AgentBench Memory Track (Tsinghua KEG, 2308)xmemory vendor documentation; evaluated on MemoryBank: Enhancing LLMs with Long-Term Memory (Sun Yat-sen University, 2305)