Siren MCP

Your agent writes the code. Now it ships the marketing.

Siren is a remote MCP server. Add one URL to Claude Code, Cursor, ChatGPT, Claude or Grok, approve once, and your agent can make an on-brand video or image, write the caption and post it. You never leave the chat.

mcp.mysiren.ai/mcp
All agentsconnected
> make a launch video for today’s release
βœ“ render_asset  launch video, 18s
> post it to X and LinkedIn
βœ“ post_run      posted

set up

Connect in three steps

  1. 1

    Add the URL as a remote MCP server in your agent. Each agent has its own page below.

  2. 2

    Approve once at app.mysiren.ai/connect. It is OAuth, so there is no API key to paste.

  3. 3

    Turn on Allow posting on the consent screen when you want the agent to publish. It stays off until you do.

https://mcp.mysiren.ai/mcp

What your agent can do with it

  • Pick the right film: browse the film library with list_films and read any film's contract with describe_film before it renders.
  • Make a video: motion graphics for a launch, a feature or a changelog, in your colors, type and logo. 25 credits, or 50 for a platinum video.
  • Make a still: one painted image from your prompt and Brand DNA, sized for the channel. 10 credits.
  • Write the caption for each platform.
  • Post now with post_run, or queue it with schedule_post at a time and timezone you pick.
  • Make it for any brand: pass a brand to create_campaign or render_asset for a client or side project, with nothing from your own workspace leaking in.
  • Read and update your Brand DNA, and upload product screenshots the next render can use.

Why an MCP server and not another dashboard

If you build with an agent, your product lives in the editor. Marketing lives somewhere else: a design tool, a scheduler, a tab you forget to open. So the feature ships and nobody hears about it.

With Siren on MCP the ask is one line in the same chat: post this. The agent already knows what you just shipped. Siren already knows what your brand looks like. The post goes out and you go back to building.

What stays in your control

  • Posting is off until you flip Allow posting.
  • Posts only go to channels you connected yourself in the Siren dashboard.
  • Every render is saved in your Siren workspace, so you can look before anything goes out.
  • You can revoke the agent any time under Developers.

The tools

Films: list_films browses the film library by job, describe_film returns a film's full contract, and list_platinum_films lists the platinum tier. list_styles browses the still styles.

Generate and post: get_account, clarify_brief, create_campaign, render_asset, list_runs, get_run, post_run, schedule_post.

Brand DNA: get_brand_dna, update_brand_dna, upload_logo. Product screens: list_product_screens, upload_product_screen, update_product_screen, delete_product_screen.

People and mascot: list_people, add_person, delete_person, get_mascot, create_mascot, mascot_step. Media: upload_media, list_media, delete_media, upload_brand_video, list_brand_videos, delete_brand_video. Results and credits: get_performance, list_posts, buy_credits (a checkout link the human pays).

create_campaign takes an optional scheduled_at, timezone and platforms (x, linkedin, tiktok, youtube, or everything). The full list with arguments is public at mcp.mysiren.ai/tools. Siren is listed in the MCP registry as ai.mysiren/siren.

faq

Straight answers.

Which plans include the MCP server?

Every plan, Spark included. Spark ($19, 100 credits) reaches the agent door the same as Pro ($49, 300 credits) and Scale ($149, 1,200 credits). Videos are the one thing that needs Pro or above.

Do I need an API key?

No. The server uses OAuth. You approve once in the browser and the agent is connected.

Can the agent post without asking me?

Only if you turn on Allow posting. Until then it can make videos and images but it cannot publish them.

Which agents work with it?

Any client that supports remote MCP over HTTP. There are setup pages for Claude Code, Cursor, ChatGPT and Claude, and Grok works with grok mcp add.

Is the video AI generated?

Videos are motion graphics rendered from code, frame by frame, so the same prompt gives the same video. Stills are painted by an image model from your prompt and Brand DNA. Voiceover is only added when you ask for it.

siren mcp

Pick your agent.

a thousand voices, one signature

Type post this. Siren does the rest.

The MCP server is on every plan, Spark included. Pro adds videos, 300 credits a month and auto-posting.

Β© 2026 Siren Ai, Inc.