API · MCP · skill

Run Svmmon from anywhere.

Drive the whole slideshow workflow from your terminal or any agent. Generate scored hooks, write slides in your account's voice, render finished slideshows, and deliver them to TikTok or Telegram, all over a simple REST API.

Get started

01 Set up

One step. Hand it to your agent.

Every one of these tools is an agent that can configure itself. Generate a key, paste the prompt below into your agent's chat, and it adds the Svmmon MCP server, installs the skill, and asks for your key. You never have to find a config file.

Generate an API key

In the Svmmon app, open Settings → API Keys and generate a key. It starts with svm_. One key covers the skill and the MCP server. No account yet? Pick a plan first. Every paid plan includes API access.

Paste this prompt to your agent

Paste it into your agent's chat (Claude Code, Codex, Cursor, Gemini, or any agent). It sets itself up: adds the MCP server and installs the skill, then asks for your key.

Agent prompt · paste into chat

Set up Svmmon for me. 1) Add an MCP server named "svmmon" that runs `npx -y svmmon-mcp` with the environment variable SVMMON_API_KEY=svm_YOUR_KEY, in whatever MCP config is correct for you (Claude Code: `claude mcp add`; Cursor: mcp.json; Codex: config.toml; Gemini: settings.json). 2) Download the Svmmon skill from https://app.svmmonapp.com/svmmon-claude-code-skill.zip and install it if you support skills, otherwise read its files as context. 3) Show me your MCP config so I can confirm, then help me make a slideshow.

Swap svm_YOUR_KEY for your key (Settings → API Keys). Works in any agent, now or later.

However you connect, the REST API works from anything that can make an HTTP request. Full spec at /openapi.json.

Prefer to set it up manually? (Claude Code, Codex, Cursor, Gemini configs)

Generate an API key

In the Svmmon app, open Settings → API Keys and generate a key. It starts with svm_. You only need one key for the skill and the MCP server.

Add Svmmon to Claude Code

One line adds the MCP server. Paste it into your terminal. Swap svm_your_key for the key you just generated.

Terminal · add the MCP server

claude mcp add svmmon -e SVMMON_API_KEY=svm_your_key -- npx -y svmmon-mcp

Or install the skill instead

Terminal · install the skill

curl -fsSL https://app.svmmonapp.com/svmmon-claude-code-skill.zip -o /tmp/svmmon-skill.zip && unzip -oq /tmp/svmmon-skill.zip -d ~/.claude/skills && rm /tmp/svmmon-skill.zip

Ask your agent

Start a new Claude Code session and ask in plain language, for example: "Make three slideshows for my Marcus profile and post them to TikTok." It makes the API calls, and everything generates in your account's persona voice.

However you connect, the REST API works from anything that can make an HTTP request. Full spec at /openapi.json.

Generate an API key

In the Svmmon app, open Settings → API Keys and generate a key. It starts with svm_.

Add Svmmon to Codex

Add the server from the CLI. Swap svm_your_key for your key.

Terminal · add the MCP server

codex mcp add svmmon --env SVMMON_API_KEY=svm_your_key -- npx -y svmmon-mcp

Or edit the config file directly

~/.codex/config.toml

[mcp_servers.svmmon]
command = "npx"
args = ["-y", "svmmon-mcp"]

[mcp_servers.svmmon.env]
SVMMON_API_KEY = "svm_your_key"

Ask your agent

In a Codex session, ask in plain language, for example: "Make three slideshows for my Marcus profile and post them to TikTok." It makes the API calls, and everything generates in your account's persona voice.

However you connect, the REST API works from anything that can make an HTTP request. Full spec at /openapi.json.

Generate an API key

In the Svmmon app, open Settings → API Keys and generate a key. It starts with svm_.
2
Add Svmmon to Cursor

Cursor is file-based: add it via Settings → MCP, or edit .cursor/mcp.json in your project (or ~/.cursor/mcp.json to make it global). Swap svm_your_key for your key.
.cursor/mcp.json
```
{
  "mcpServers": {
    "svmmon": {
      "command": "npx",
      "args": ["-y", "svmmon-mcp"],
      "env": {
        "SVMMON_API_KEY": "svm_your_key"
      }
    }
  }
}
```
Ask your agent

Open Cursor's chat and ask in plain language, for example: "Make three slideshows for my Marcus profile and post them to TikTok." It makes the API calls, and everything generates in your account's persona voice.

However you connect, the REST API works from anything that can make an HTTP request. Full spec at /openapi.json.

Generate an API key

In the Svmmon app, open Settings → API Keys and generate a key. It starts with svm_.

Add Svmmon to Gemini CLI

Add the server from the CLI. Swap svm_your_key for your key.

Terminal · add the MCP server

gemini mcp add svmmon -e SVMMON_API_KEY=svm_your_key -- npx -y svmmon-mcp

Or edit the config file directly

~/.gemini/settings.json

{
  "mcpServers": {
    "svmmon": {
      "command": "npx",
      "args": ["-y", "svmmon-mcp"],
      "env": {
        "SVMMON_API_KEY": "svm_your_key"
      }
    }
  }
}

Ask your agent

In a Gemini CLI session, ask in plain language, for example: "Make three slideshows for my Marcus profile and post them to TikTok." It makes the API calls, and everything generates in your account's persona voice.

However you connect, the REST API works from anything that can make an HTTP request. Full spec at /openapi.json.

Generate an API key

In the Svmmon app, open Settings → API Keys and generate a key. It starts with svm_.
2
Add Svmmon

This covers everything else: Windsurf, Cline, Copilot in VS Code, Claude Desktop, Zed, and any other MCP-compatible tool. Add an MCP server however your tool does it, pointing at npx -y svmmon-mcp with your SVMMON_API_KEY. The Svmmon side is identical everywhere. Most tools take the JSON block below verbatim.
mcpServers config
```
{
  "mcpServers": {
    "svmmon": {
      "command": "npx",
      "args": ["-y", "svmmon-mcp"],
      "env": {
        "SVMMON_API_KEY": "svm_your_key"
      }
    }
  }
}
```
If your tool cannot do MCP, call the REST API directly. It is the universal floor and works from any language or script.
Ask your agent

Once the server is connected, ask in plain language, for example: "Make three slideshows for my Marcus profile and post them to TikTok." It makes the API calls, and everything generates in your account's persona voice.

However you connect, the REST API works from anything that can make an HTTP request. Full spec at /openapi.json.

02 What the skill is

A self-contained knowledge pack for the Svmmon API.

The skill is a folder of plain files that document the whole Svmmon REST API. Claude Code loads it as a skill, so you can ask for slideshows in your account's voice and it knows exactly which calls to make. Other agents can read the same files as context, or just point at the OpenAPI spec, since every capability is in the API.

What is in the box

SKILL.md tells the agent when and how to use Svmmon.
api-reference.md documents every endpoint and field.
faq.md covers the common questions and edge cases.
openapi.json is the machine-readable spec for the whole API.

Persona-driven, end to end

Your account already holds the voice, audience, and style of each persona you have set up. Generation runs in that voice, so slideshows come back sounding like you, not like a generic model.

Ask in plain language. On Claude Code the skill handles the calls; on any other agent the MCP server gives you the same persona-driven generation.

03 What you can do

The endpoints, at a glance.

Authenticate with a Bearer svm_ key. Base URL is https://app.svmmonapp.com/api/v1. Sixteen endpoints cover the full workflow, from reading what works to posting the result.

Read

Check your usage and remaining quota.
List your profiles (your personas).
List the slide presets you can generate against.
List your past slideshows and image collections.
Pull post performance and TikTok insights.
Read your brain: what is working across your accounts.
Browse studio history, providers, and saves.

Create and deliver

Generate scored hooks for a profile.
Generate full, rendered slideshows.
Deliver a slideshow to TikTok.
Deliver a slideshow to Telegram.

Terminal · generate five scored hooks

curl https://app.svmmonapp.com/api/v1/hooks/generate \
  -H "Authorization: Bearer svm_your_key" \
  -H "Content-Type: application/json" \
  -d '{"profile_id": "uuid", "count": 5}'

Response · trimmed

{
  "hooks": [
    { "text": "I posted the same slideshow 3 ways and one did 40x the views", "score": 78 },
    { "text": "nobody warns you the first 200 followers are the slowest part", "score": 71 },
    { "text": "the posting time myth cost me a month of growth", "score": 64 }
  ]
}

Profile IDs come from GET /api/v1/profiles. Scores use the same 0-100 virality scorer as the app.

Machine-readable spec

The full OpenAPI spec is public. Point any tool, generator, or agent at it.

app.svmmonapp.com/openapi.json

04 Access

Available on every paid plan.

The API and the skill come with every paid plan, and that includes TikTok insights. Generation counts against your plan's monthly volume. Full limits are on the pricing table. You generate your API key in the app right after you sign up.

The short version. Sign up, pick any paid plan, generate an API key, then install the skill or add the MCP server. After that, you drive Svmmon from Claude Code, Cursor, Codex, or any agent.

Get started See pricing

One step. Hand it to your agent.

Generate an API key

Paste this prompt to your agent

Generate an API key

Add Svmmon to Claude Code

Ask your agent

Generate an API key

Add Svmmon to Codex

Ask your agent

Generate an API key

Add Svmmon to Cursor

Ask your agent

Generate an API key

Add Svmmon to Gemini CLI

Ask your agent

Generate an API key

Add Svmmon

Ask your agent

A self-contained knowledge pack for the Svmmon API.

What is in the box

Persona-driven, end to end

The endpoints, at a glance.

Read

Create and deliver

Machine-readable spec

Available on every paid plan.