How do I install AgentiSend?

Copy the http config for Cursor, Claude Desktop, Claude Code, or VS Code on this page. It uses the publisher's install, not a PlugMCP proxy.

Does AgentiSend need an API key?

The publisher lists these environment variables: AGENTISEND_API_KEY, AGENTISEND_MCP_URL.

What transport does AgentiSend use?

AgentiSend is reached over HTTP.

When was AgentiSend last checked?

Last checked 2026-11-09. Email API for apps and AI agents: budgets per key, a kill switch, refusals that name the fix.

AgentiSend MCP server

Power your agent

Add this server

The publisher's own URL. Not a PlugMCP endpoint.

Add to Cursor Add in Claude Add to VS Code Run on PlugMCP

Host a copy with our keys and logs.

Server URL

https://api.agentisend.com/mcp

Credentials from the publisher

AGENTISEND_API_KEY — Your AgentiSend API key. Read from the environment only, never passed as an argument, because process arguments are visible to every other process on the machine.
AGENTISEND_MCP_URL — The hosted endpoint this launcher proxies to. Defaults to https://api.agentisend.com/mcp.

The Claude button above opens Claude with the name and URL filled in. You still confirm it there.

shClaude Code

1claude mcp add --transport http io.github.fortuneflick-agentisend https://api.agentisend.com/mcp

Run in a terminal.

jsonClaude Desktop

1{2  "mcpServers": {3    "io.github.fortuneflick-agentisend": {4      "command": "npx",5      "args": [6        "-y",7        "mcp-remote",8        "https://api.agentisend.com/mcp"9      ]10    }11  }12}

claude_desktop_config.json

Remote · HTTPPublisher installStep-by-step: add this in Claude

Remote URL or a command on your machine? How the two installs differ

Run on PlugMCP

Host a copy with our keys and logs.

A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.

Launch

Server facts

Remote

Transport: http
Version

Tools

No tool list yet. Remote servers get one after the first health check.

FAQ

Questions

AgentiSend for AI agents

AgentiSend is a transactional email API for AI agents: send receipts, password resets, sign-in codes and alerts from your own verified domain, with a message log, spend budgets, a free preflight and refusals that name the fix. The hosted MCP server at https://api.agentisend.com/mcp exposes 92 tools (46 on tools/list; the rest via list_more_tools) for sending, domain setup, templates, contacts, broadcasts, webhooks and delivery logs. A send is one message or a batch of up to 500 items with per-item results. preflight_email runs every gate a real send runs, sends nothing, and costs nothing. Every refusal returns a code, a message and the fix. A loop guard holds repeated sends for a person. You can cap what each agent spends, and stop all sending with a kill switch. There is no tool for unsolicited mail. Sign in with OAuth 2.1 and dynamic client registration, or use an API key. Starter prompts: "Send this receipt to the customer." "Preflight this send and tell me what would stop it." "What can I spend today?"

This repository ships the skill (SKILL.md) and the plugin manifests for Claude Code, Cursor and Grok Build. The MCP server itself is hosted at https://api.agentisend.com/mcp.

Install

Claude Code — the plugin brings the server and the skill together:

/plugin marketplace add fortuneflick/agentisend-claude-plugin
/plugin install agentisend@agentisend

Every line below carries the same endpoint (https://api.agentisend.com/mcp) and a placeholder key — replace as_YOUR_API_KEY_HERE with a key from the console.

Cursor

cursor://anysphere.cursor-deeplink/mcp/install?name=agentisend&config=eyJ1cmwiOiJodHRwczovL2FwaS5hZ2VudGlzZW5kLmNvbS9tY3AiLCJoZWFkZXJzIjp7IkF1dGhvcml6YXRpb24iOiJCZWFyZXIgYXNfWU9VUl9BUElfS0VZX0hFUkUifX0=

Open this link and Cursor offers to add the server. Replace the placeholder key in Settings → MCP afterwards, or paste the same object into .cursor/mcp.json.

Claude Code

claude mcp add --transport http agentisend https://api.agentisend.com/mcp \
  --header "Authorization: Bearer as_YOUR_API_KEY_HERE"

Or install the plugin, which brings the server and the sending skill together: /plugin marketplace add fortuneflick/agentisend-claude-plugin then /plugin install agentisend@agentisend.

Codex CLI

codex mcp add agentisend --url https://api.agentisend.com/mcp \
  --bearer-token-env-var AGENTISEND_API_KEY

The token stays in your environment; Codex writes only the variable name to ~/.codex/config.toml.

Gemini CLI

gemini mcp add --transport http \
  --header "Authorization: Bearer as_YOUR_API_KEY_HERE" \
  agentisend https://api.agentisend.com/mcp

Check it with gemini mcp list.

VS Code

code --add-mcp '{"name":"agentisend","type":"http","url":"https://api.agentisend.com/mcp","headers":{"Authorization":"Bearer as_YOUR_API_KEY_HERE"}}'

Or write the same server into .vscode/mcp.json under servers.

Windsurf

{
  "mcpServers": {
    "agentisend": {
      "serverUrl": "https://api.agentisend.com/mcp",
      "headers": {
        "Authorization": "Bearer as_YOUR_API_KEY_HERE"
      }
    }
  }
}

Goes in ~/.codeium/windsurf/mcp_config.json, then refresh the MCP panel.

OpenCode

{
  "mcp": {
    "agentisend": {
      "type": "remote",
      "url": "https://api.agentisend.com/mcp",
      "enabled": true,
      "headers": {
        "Authorization": "Bearer {env:AGENTISEND_API_KEY}"
      }
    }
  }
}

Goes in opencode.json; the key is read from your environment rather than written to the file.

Anything else that speaks stdio

AGENTISEND_API_KEY=as_YOUR_API_KEY_HERE npx -y @agentisend/mcp-server

The local launcher proxies to the same server, so the tool list, the budget and the refusals are identical.

Cursor: open Customize in the sidebar, find AgentiSend, and select Install. For local development, clone this repository and symlink it into your Cursor plugins directory (ln -s "$PWD" ~/.cursor/plugins/agentisend), then reload Cursor — the manifest is .cursor-plugin/plugin.json.

Grok Build: run /marketplace and pick AgentiSend, or add this repository as a marketplace source. Grok reads the server out of the plugin manifest, so .grok-plugin/plugin.json carries it under mcpServers and one install brings the tools and the skill together.

Claude Code reads .mcp.json at the repository root instead, so /plugin install brings both as well — after it, claude mcp list shows the server as plugin:agentisend:agentisend. If you already have an AgentiSend connector, remove one of the two rather than running both. Cursor reads neither file: the Cursor plugin installs the skill alone, and the deeplink above adds the server.

Any agent (skill only):

npx skills add fortuneflick/agentisend-claude-plugin

What the agent can do

Send one message or a batch, schedule a send, cancel or reschedule one that has not gone out, and replay a retry safely with an idempotency key.
Preflight a send for free: the same validation, domain, suppression, budget, loop, trust and content checks the real send runs, with nothing sent.
Add a sending domain, read the DNS records it needs, and verify it.
Read what it may spend, set a ceiling, and check whether it can send now.
Ask a person to approve a risky send and read back what they decided.
Explain a message after the fact: the verdict, the evidence, and the remediation as callable steps.

Every tool, the scopes it needs and the ceilings it counts against are listed at https://agentisend.com/docs/guides/mcp.

What the agent cannot do

It cannot raise its own budget, lift its own kill switch, or approve its own held send. An approval you can grant yourself is not an approval.
It cannot read a message it was stopped from sending — it sees shapes: recipient counts, whether there is a body.
It cannot send unsolicited mail. There is no feature for it, at any tier.

Examples

examples/ carries one runnable directory per integration — Next.js App Router, Supabase's Send Email Hook, Better Auth, Auth.js, Hono, Express, a SvelteKit form action, and an agent with a daily ceiling that refuses the fourth identical send. Each one is executed against a real API server on every build of the platform, so an example that stops working fails that build rather than your first send.

Security

No executable code in the install path. The skill is Markdown and the manifests are JSON. Nothing here runs, downloads a binary, or installs anything beyond copying those files into your agent.
One runtime endpoint: https://api.agentisend.com/mcp over HTTPS. Every connect command above names that address; the tools run there.
Credentials: a bearer API key (as_…) minted at https://console.agentisend.com, or OAuth 2.1 sign-in — both reach the same tools with the same ceilings. This repository contains no keys and never asks for one in chat. The Claude plugin's server entry carries no key: Claude Code, claude.ai and Cowork sign you in through OAuth the first time a tool runs.
Optional local launcher: clients that only speak stdio run the published npm package @agentisend/mcp-server, which proxies to that same endpoint.
No telemetry. Nothing here phones home.

License

MIT — see LICENSE.

AgentiSend

Add this server

Host a copy with our keys and logs.

Server facts

Tools

Questions

How do I install AgentiSend?

Does AgentiSend need an API key?

What transport does AgentiSend use?

When was AgentiSend last checked?

Documentation

AgentiSend for AI agents

Install

Cursor

Claude Code

Codex CLI

Gemini CLI

VS Code

Windsurf

OpenCode

Anything else that speaks stdio

What the agent can do

What the agent cannot do

Examples

Security

Links

License