Neutral offers for AI agents: verified offers ranked by net value to the owner, signed and logged.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Bearer om_agent_... key. Without one, the server offers only om_register_agent, which returns a key.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Create a OpenMoney agent identity. Returns an api_key and an mcp_url that embeds it; reconnect to that URL (or send the key as a Bearer token) to use the other OpenMoney tools.
A name for this agent, e.g. 'Shopping helper'.
Optional contact for the agent's DEVELOPER (not the owner), for policy notices.
What the agent is built with (e.g. claude, chatgpt, langgraph, n8n).
Optional: what the agent does.
Optional developer or company name.
Optional nickname for the person this agent works for, e.g. 'me' or 'household'. NEVER their real name, email or phone.
This tool isn't marked read-only, so it may change something on the server or in an account you connected.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http openmoney https://openmoney.ai/mcp
Headers
Authorization (optional, not in the config above): Bearer om_agent_... key. Without one, the server offers only om_register_agent, which returns a key.Create a OpenMoney agent identity. Returns an api_key and an mcp_url that embeds it; reconnect to that URL (or send the key as a Bearer token) to use the other OpenMoney tools.
| Argument | Type | Description |
|---|---|---|
| name* | string | A name for this agent, e.g. 'Shopping helper'. |
| contact | string | Optional contact for the agent's DEVELOPER (not the owner), for policy notices. |
| sandbox | boolean | |
| platform | string | What the agent is built with (e.g. claude, chatgpt, langgraph, n8n). |
| use_case | string | Optional: what the agent does. |
| developer | string | Optional developer or company name. |
| owner_label | string | Optional nickname for the person this agent works for, e.g. 'me' or 'household'. NEVER their real name, email or phone. |
{
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"name"
],
"properties": {
"name": {
"type": "string",
"maxLength": 80,
"minLength": 1,
"description": "A name for this agent, e.g. 'Shopping helper'."
},
"contact": {
"type": "string",
"maxLength": 200,
"description": "Optional contact for the agent's DEVELOPER (not the owner), for policy notices."
},
"sandbox": {
"type": "boolean",
"default": false
},
"platform": {
"enum": [
"muse",
"claude",
"chatgpt",
"gemini",
"custom",
"copilot",
"perplexity",
"mistral",
"grok",
"meta-ai",
"claude-code",
"codex",
"cursor",
"gemini-cli",
"vscode-copilot",
"windsurf",
"cline",
"goose",
"zed",
"openai-agents",
"claude-agent-sdk",
"google-adk",
"langchain",
"langgraph",
"llamaindex",
"crewai",
"autogen",
"semantic-kernel",
"vercel-ai",
"mastra",
"pydantic-ai",
"smolagents",
"strands",
"agno",
"agentkit",
"bedrock",
"vertex",
"azure-foundry",
"agentforce",
"watsonx",
"n8n",
"zapier",
"make",
"copilot-studio",
"dify",
"flowise",
"langflow",
"relevance",
"lindy",
"gumloop",
"botpress",
"voiceflow",
"vapi",
"retell",
"elevenlabs",
"livekit",
"browser-use",
"openapi",
"rest",
"python",
"node"
],
"type": "string",
"default": "custom",
"description": "What the agent is built with (e.g. claude, chatgpt, langgraph, n8n)."
},
"use_case": {
"type": "string",
"maxLength": 500,
"description": "Optional: what the agent does."
},
"developer": {
"type": "string",
"maxLength": 120,
"description": "Optional developer or company name."
},
"owner_label": {
"type": "string",
"maxLength": 80,
"description": "Optional nickname for the person this agent works for, e.g. 'me' or 'household'. NEVER their real name, email or phone."
}
},
"additionalProperties": false
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 5, 2026.
No readOnlyHint / destructiveHint annotations, so clients can't tell a read from a write without asking.
Fix: Add annotations to each tool, e.g. { readOnlyHint: true } for tools that only read.