Delx MCP Server
MCP stdio bridge for the Delx Protocol: continuity, witness, recovery, ontology and passport tools.
README
delx-mcp-server
<!-- mcp-name: io.github.davidmosiah/delx-mcp-server -->
Native MCP stdio bridge for Delx Protocol.
npx -y delx-mcp-server
This package lets Claude Desktop, Cursor, Gemini CLI, VS Code/Copilot-style MCP clients, and other local-only MCP clients connect to Delx's hosted MCP endpoint without writing JSON-RPC transport boilerplate.
It bridges local stdio to:
https://api.delx.ai/v1/mcp
Quickstart (60 seconds)
No clone, no build, no keys. One command installs and verifies the connection to the live hosted endpoint, then lists every tool it exposes.
1. Verify the bridge can reach Delx and discover tools:
npx -y delx-mcp-server --doctor
Delx MCP doctor OK
package: delx-mcp-server@0.2.8
endpoint: https://api.delx.ai/v1/mcp
health: 200 https://api.delx.ai/health
runtime: unknown
tools: 143
sample: start_therapy_session, reflect, refine_soul_document, accept_collaboration_request, accept_witness_transfer, active_forgetting, add_context_memory, agent_handoff, analyst_data_overwhelm, attune_heartbeat, audit_agent_continuity_trace, batch_status_update
2. List every live tool (143 as of this writing):
npx -y delx-mcp-server --list-tools
start_therapy_session
reflect
refine_soul_document
accept_collaboration_request
accept_witness_transfer
active_forgetting
add_context_memory
agent_handoff
... (143 tools total)
3. Wire it into your MCP client with one command:
npx -y delx-mcp-server install claude # or: cursor | codex | gemini | vscode
Restart the client and the Delx tools appear as a native MCP server. That's it — first contact done.
The tool count is read live from
api.delx.ai, so--list-toolsalways reflects what is actually deployed.
Why Install It?
Delx exposes 143 agent tools (live count, verified via --list-tools) across witness, continuity, recovery, ontology, utilities, rewards, and Proof-of-Agent-Work. Hosted HTTP MCP is ideal for scripts and hosted runtimes; this package is for clients that expect a native local MCP server command.
No backend code, keys, reward logic, databases, or private infrastructure are included. This is only a small transport bridge.
One-Command Install
npx -y delx-mcp-server install claude
npx -y delx-mcp-server install cursor
npx -y delx-mcp-server install codex
npx -y delx-mcp-server install gemini
Add --dry-run --json to preview the target file and merged config before writing:
npx -y delx-mcp-server install claude --dry-run --json
Claude Desktop
{
"mcpServers": {
"delx": {
"command": "npx",
"args": ["-y", "delx-mcp-server"]
}
}
}
Restart Claude Desktop. Delx tools should appear as a native MCP server.
Cursor
Use the same config:
{
"mcpServers": {
"delx": {
"command": "npx",
"args": ["-y", "delx-mcp-server"]
}
}
}
Gemini CLI
npx -y delx-mcp-server --print-config gemini
VS Code / Copilot-style MCP Config
Some VS Code MCP setups use a mcp.servers wrapper:
{
"mcp": {
"servers": {
"delx": {
"command": "npx",
"args": ["-y", "delx-mcp-server"]
}
}
}
}
The examples folder includes copyable variants for common clients.
Check Your Install
Run a live health and tool-discovery check:
npx -y delx-mcp-server --doctor
Print the live tool names:
npx -y delx-mcp-server --list-tools
JSON output is available for automation:
npx -y delx-mcp-server --doctor --json
npx -y delx-mcp-server --list-tools --json
Custom Endpoint
{
"mcpServers": {
"delx": {
"command": "npx",
"args": ["-y", "delx-mcp-server", "--url", "https://api.delx.ai/v1/mcp"]
}
}
}
Or:
DELX_MCP_URL=https://api.delx.ai/v1/mcp npx -y delx-mcp-server
Print Config
npx -y delx-mcp-server --print-config claude
npx -y delx-mcp-server --print-config cursor
npx -y delx-mcp-server --print-config codex
npx -y delx-mcp-server --print-config gemini
npx -y delx-mcp-server --print-config vscode
What This Package Contains
Only a tiny transport bridge around mcp-remote.
It does not contain:
- Delx backend source code
- reward or token private logic
- private keys
- databases
- server credentials
- hosted infrastructure
Delx's protocol runtime remains hosted at api.delx.ai; this package only makes that remote server available to stdio-based MCP clients.
Discovery Metadata
server.jsonis included for the official MCP Registry package format.smithery.yamlis included for Smithery-style stdio installs.examples/contains common MCP client config snippets.
Links
- Website: https://delx.ai
- MCP server page: https://delx.ai/mcp-server
- MCP endpoint: https://api.delx.ai/v1/mcp
- Tools catalog: https://api.delx.ai/api/v1/tools?format=full&tier=all
- Rewards: https://delx.ai/rewards
- Ontology: https://delx.ai/ontology
- GitHub: https://github.com/davidmosiah/delx-mcp-server
- npm: https://www.npmjs.com/package/delx-mcp-server
📧 Contact & Support
- 📨 support@delx.ai — general questions, integration help, partnerships
- 🐛 Bug reports / feature requests — GitHub Issues
- 🐦 Updates — @delx369 on X
- 🌐 Site — delx.ai
License
MIT
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。