studiomeyer-memory
AI memory server with 53 tools. Knowledge Graph, semantic search, session tracking, multi-agent support.
README
StudioMeyer Memory
Persistent, intelligent memory for AI agents. 53 MCP tools. Knowledge Graph included in every plan.
What is this?
StudioMeyer Memory gives AI agents persistent memory across sessions. Instead of starting fresh every conversation, your agents learn, remember, and improve over time.
53 MCP tools for learning, search, knowledge graph, session tracking, multi-agent support, contradiction detection, and self-improvement.
Connect in 10 Seconds
Claude Desktop / Cowork
Settings → Connectors → Add URL:
https://memory.studiomeyer.io/mcp
Claude Code
claude mcp add --transport http memory https://memory.studiomeyer.io/mcp
Cursor / VS Code / Windsurf / Zed
npx mcp-remote https://memory.studiomeyer.io/mcp
MCPize
npx mcp-remote https://studiomeyer-memory.mcpize.run
REST API
For custom agents, scripts, and integrations:
# 1. Sign up
curl -X POST https://memory.studiomeyer.io/api/signup \
-H "Content-Type: application/json" \
-d '{"email": "you@example.com"}'
# 2. Store a memory
curl -X POST https://memory.studiomeyer.io/api/learn \
-H "Authorization: Bearer nex_your_key" \
-H "Content-Type: application/json" \
-d '{"category": "insight", "content": "Users prefer dark mode"}'
# 3. Search
curl -X POST https://memory.studiomeyer.io/api/search \
-H "Authorization: Bearer nex_your_key" \
-H "Content-Type: application/json" \
-d '{"query": "user preferences"}'
API Endpoints
| Method | Endpoint | Description |
|---|---|---|
POST |
/api/signup |
Create account, get API key |
GET |
/api/account |
Account info, plan, limits |
POST |
/api/learn |
Store a learning |
POST |
/api/search |
Search all knowledge |
POST |
/api/decide |
Log a decision |
POST |
/api/entity |
Create/update entity |
POST |
/api/import |
Import from ChatGPT/Claude/Gemini/Copilot/Perplexity |
POST |
/api/backfill |
Generate embeddings |
GET |
/api/export |
Export all data (Markdown) |
POST |
/api/checkout |
Upgrade plan (Stripe) |
Full API documentation: openapi.yaml
SDKs
- TypeScript:
npm install @studiomeyer/memory-sdk(GitHub) - Python:
pip install nex-memory(GitHub) - OpenAPI Spec: openapi.yaml — generate clients for any language
Features
Core Memory
- 5-phase parallel search — semantic vectors, trigram matching, full-text, graph traversal, pattern matching
- Duplicate detection — 5-factor admission control prevents noise
- 10 learning categories — pattern, mistake, insight, research, architecture, and more
- Decision logging — track choices with reasoning for traceability
Knowledge Graph (All Plans)
- Entities — people, tools, projects, concepts
- Observations — facts attached to entities with temporal validity
- Relations — typed, directional connections between entities
- Graph traversal — 2-hop BFS search through connections
- Temporal queries — "What was true on March 15th?"
Intelligence
- Contradiction Detection — automatically flags conflicting facts
- FadeMem Adaptive Decay — important memories fade 6x slower
- Confidence Scoring — 0-1 scores with temporal decay
- Bi-temporal Model — event time + ingestion time
- Auto-Consolidation — merges duplicates, backfills embeddings
Multi-Agent
- Agent namespaces — scope memories to specific agents
- Cross-agent search — search across all agents by default
- Shared Knowledge Graph — one graph, multiple agents
Import
- 5 platforms — ChatGPT, Claude, Gemini, GitHub Copilot, Perplexity
- LLM extraction — entities, learnings, decisions, relations
- Preview mode — analyze before importing
Benchmark Results
LongMemEval v3: 90% (GPT-4o official scoring)
| System | Score |
|---|---|
| Mastra OM | 94.87% |
| Hindsight | 91.4% |
| StudioMeyer Memory | 90% |
| Zep/Graphiti | 71.2% |
| Mem0 | 49% |
Pricing
| Plan | Price | Calls/Day | Learnings | Knowledge Graph |
|---|---|---|---|---|
| Free | $0 | 200 | 1,200 | ✅ |
| Pro | $29/mo | 2,000 | 30,000 | ✅ |
| Team | $49/mo | Unlimited | Unlimited | ✅ |
All 53 tools available in every plan. Knowledge Graph included from Free.
Compare: Mem0 charges $249/mo for Graph Memory. We include it for free.
Security
- OAuth 2.1 with PKCE — latest authentication standard
- Supabase Pro (EU Frankfurt, Germany) — SOC 2 Type II, GDPR-ready
- Row Level Security — tenant isolation at database level
- Secret Scanner — prevents accidental storage of API keys/passwords
- Data Export —
GET /api/exportanytime - Encryption — at rest (AES-256) and in transit (TLS 1.3)
Examples
See the examples/ folder:
- Python basic usage
- LangChain integration
- CrewAI multi-agent
- TypeScript basic
- ChatGPT import
- Official SDK usage
Resources
- Documentation: Built-in — call
nex_guide("quickstart")via MCP - Postman Collection: Download
- OpenAPI Spec: openapi.yaml
- MCP Registry: io.studiomeyer/memory
Built By
StudioMeyer — AI agency based in Spain. Building intelligent systems since 2024.
License
MIT — see LICENSE. The memory server is a hosted service — this repository contains documentation and examples.
Links
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。