Firmament Open Brain
Firmament Open Brain
为 Firmament 团队打造的一个共享“上下文大脑”,灵感来自 OB1,但为 Railway 进行了精简重建。它是一个以 Postgres 为后端的 MCP(Model Context Protocol)服务器:每个人都有一个私有记忆库,可以在 AI 工具(Claude Desktop、Claude Code、claude.ai、Cursor 等)之间跟随使用。记忆按用户隔离——你的 token 永远只能看到你自己的记忆。
工作原理
一个 Railway 服务(这个 Node 应用)+ 一个 Railway Postgres。
每个用户有一个个人 token。他们的 MCP URL 为
https://<host>/mcp/<token>。所有记忆读写都限定于该 token 解析到的用户。Token 以 SHA-256 哈希存储。
搜索采用 Postgres 全文搜索,并提供子字符串回退——无需 embedding API 密钥。
Related MCP server: LedgerMem MCP Server
暴露的 MCP 工具
工具 | 功能 |
| 存储一条记忆(内容 + 可选标签) |
| 按主题/关键词搜索你的记忆 |
| 列出最新记忆 |
| 按 id 删除一条记忆 |
连接客户端
Claude Code:
claude mcp add --transport http openbrain https://<host>/mcp/<your-token>**Claude Desktop / claude.ai:**设置 → 连接器 → 添加自定义连接器 → 粘贴 https://<host>/mcp/<your-token>。
**Cursor:**设置 → MCP → 添加服务器,类型选 http,使用相同 URL。
然后直接说:“记住……” / “关于……你知道些什么?”一个好习惯是让会话以“检查我的 open brain 以获取关于 的上下文”开头。
管理员(创建用户)
需要 ADMIN_TOKEN 环境变量;将其作为 Bearer token 传递。
# create a user (returns their token + MCP URL — shown only once)
curl -X POST https://<host>/admin/users \
-H "Authorization: Bearer $ADMIN_TOKEN" \
-H "Content-Type: application/json" \
-d '{"name": "Jack Wright"}'
# list users
curl https://<host>/admin/users -H "Authorization: Bearer $ADMIN_TOKEN"
# delete a user (and all their memories)
curl -X DELETE https://<host>/admin/users/<id> -H "Authorization: Bearer $ADMIN_TOKEN"部署(Railway)
应用服务上的环境变量:
DATABASE_URL— 引用 Railway PostgresADMIN_TOKEN— 长随机字符串(管理员 API 认证)DATABASE_SSL— 如果通过 Railway 私有网络连接,则设为false
在此目录中运行 railway up 进行部署。Schema 会在启动时自动创建。
安全说明
Token 位于 URL 中。这正是连接只需一键粘贴的原因,对于小型可信团队来说可以接受,但要像对待密码一样对待该 URL:不要分享它,不要粘贴到群聊中。
丢失 token → 删除该用户,再生成一个新的(记忆与用户行绑定,因此删除用户会删除其记忆——如有需要,请先导出)。
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceShared memory for any MCP-compatible AI, enabling cross-platform knowledge persistence and retrieval so users' context, preferences, and project info follow them everywhere.3AGPL 3.0

LedgerMem MCP Serverofficial
AlicenseNot gradedqualityBmaintenanceEnables persistent memory storage and retrieval for MCP clients, allowing AI assistants to remember facts and context across conversations.10MIT- AlicenseNot gradedqualityCmaintenanceProvides persistent, local-first AI memory across sessions via MCP tools for storing, searching, and retrieving context from past interactions.1MIT
- AlicenseNot gradedqualityBmaintenanceProvides long-term memory for AI agents via MCP tools to store, recall, and delete memories, with per-user scoping and usage limits.AGPL 3.0
Related MCP Connectors
Shared, governed long-term memory for AI agents across tools and sessions via MCP and REST.
Your memory, everywhere AI goes. Build knowledge once, access it via MCP anywhere.
Person-owned AI memory that learns, not just stores — portable context for any MCP client.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/tswic-highdesert/openbrain-railway'
If you have feedback or need assistance with the MCP directory API, please join our Discord server