Marvenn MCP Server
The Marvenn MCP Server exposes the Marvenn AI-powered outreach and CRM platform to any MCP-compatible client (Claude Desktop, Continue, Cursor, etc.).
Connect your AI assistant to Marvenn and let it list campaigns, inspect leads, manage pending agent actions, and more — all through natural language.
Authentication
API key is optional for discovery (tools/list, prompts/list, resources/list).
API key is required for executing tools (tools/call).
Pass your key as a Bearer token: Authorization: Bearer
API keys are scoped. The scope granted to your key determines which tools are available:
| Scope | Description |
|---|---|
read |
Read-only access to campaigns, leads, actions, and agents |
write |
Everything in read, plus creating campaigns and approving/rejecting actions |
admin |
Everything in write, plus user management and credit operations (also requires your Marvenn account role to be admin or super_admin) |
Generate or manage your API keys from your account settings at marvenn.ai.
Transport
| Transport | URL | Status |
|---|---|---|
| Streamable HTTP (recommended) | https://marvenn.ai/mcp |
Working |
| SSE | https://marvenn.ai/mcp/sse |
Working |
Quick Start
Claude Desktop
Add the following to your claude_desktop_config.json (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):```json
{
"mcpServers": {
"marvenn": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://marvenn.ai/mcp/sse",
"--header",
"Authorization: Bearer <your-marvenn-api-key>"
]
}
}
}
Restart Claude Desktop. You should see the Marvenn tools available in the tools panel.
### curl
Open the SSE stream in one terminal:```bash
curl -N \
-H "Authorization: Bearer <your-marvenn-api-key>" \
https://marvenn.ai/mcp/sse
The server responds with an endpoint event that provides a session-specific message URL, for example:```
event: endpoint
data: https://marvenn.ai/mcp/sse/message?sessionId=mcp_1234567890_abc12345
Send a JSON-RPC message to that endpoint in a second terminal:```bash
curl -X POST \
-H "Content-Type: application/json" \
-H "Authorization: Bearer <your-marvenn-api-key>" \
-d '{"jsonrpc":"2.0","method":"tools/list","id":1}' \
"https://marvenn.ai/mcp/sse/message?sessionId=mcp_1234567890_abc12345"
The result is delivered back over the open SSE stream.
Tools
list_campaigns — Read scope
List all campaigns for the authenticated user. Returns campaign name, status, creation date, and basic stats.
| Parameter | Type | Required | Description |
|---|---|---|---|
status |
string | No | Filter by status: active, paused, completed, draft |
limit |
number | No | Max results to return (default 20, max 100) |
get_campaign_stats — Read scope
Get detailed analytics and performance stats for a specific campaign.
| Parameter | Type | Required | Description |
|---|---|---|---|
campaign_id |
number | Yes | Numeric ID of the campaign |
create_campaign — Write scope
Create a new outreach campaign.
| Parameter | Type | Required | Description |
|---|---|---|---|
name |
string | Yes | Campaign name (max 200 characters) |
type |
string | Yes | Campaign type: email, phone, social, multi_channel |
targetAudience |
string | No | Description of the target audience |
objective |
string | No | Campaign objective or goal |
list_leads — Read scope
List leads in the CRM with optional filtering and pagination.
| Parameter | Type | Required | Description |
|---|---|---|---|
status |
string | No | Filter by lead status: new, contacted, qualified, converted, lost |
search |
string | No | Search term matched against name, email, or company |
limit |
number | No | Max results per page (default 20, max 100) |
page |
number | No | Page number, 1-indexed (default 1) |
list_pending_actions — Read scope
List all pending actions awaiting your approval.
| Parameter | Type | Required | Description |
|---|---|---|---|
limit |
number | No | Max results to return (default 20) |
approve_action — Write scope
Approve or reject a pending action that requires human authorization.
| Parameter | Type | Required | Description |
|---|---|---|---|
action_id |
number | Yes | Numeric ID of the pending action |
decision |
string | Yes | approve or reject |
notes |
string | No | Optional notes or reason for the decision |
list_agents — Read scope
List all AI agents for the authenticated user.
| Parameter | Type | Required | Description |
|---|---|---|---|
limit |
number | No | Max results to return (default 20) |
list_users — Admin scope
List all users on the platform. Requires admin scope and an admin account role.
| Parameter | Type | Required | Description |
|---|---|---|---|
limit |
number | No | Max results to return (default 50, max 500) |
subscriptionTier |
string | No | Filter by tier: free, startup, team |
manage_credits — Admin scope
View or adjust the credit balance for any user. Requires admin scope and an admin account role.
| Parameter | Type | Required | Description |
|---|---|---|---|
target_user_id |
string | Yes | ID of the user whose credits to manage |
action |
string | Yes | get, add, or subtract |
amount |
number | Required for add/subtract |
Number of credits to add or subtract |
reason |
string | Required for add/subtract |
Reason for the adjustment |
Error Handling
All errors are returned as JSON-RPC error objects with the following codes:
| Code | Meaning |
|---|---|
-32600 |
Invalid JSON-RPC request |
-32601 |
Method or tool not found |
-32602 |
Invalid arguments |
-32603 |
Internal server error |
Scope errors (e.g. calling create_campaign with a read-only key) are returned as a tool result with an error field in the response body, not as a JSON-RPC protocol error.
Links
- Website: https://marvenn.ai
- MCP Protocol: https://modelcontextprotocol.io
