video.future/future-video-studio

video.future/future-video-studio

Enables agents to create, monitor, cancel, and download cinematic AI video renders through the Future Video Studio Agent API.

Category
Visit Server

README

Future Video Studio MCP

<!-- mcp-name: video.future/future-video-studio -->

Future Video Studio MCP lets agents create, monitor, cancel, and download cinematic AI video renders through the Future Video Studio Agent API.

The hosted MCP server is available at:

https://mcp.future.video/mcp

The same server can also run locally over stdio from this Python package.

The hosted service also serves the ChatGPT Apps domain verification endpoint at https://mcp.future.video/.well-known/openai-apps-challenge. Set OPENAI_APPS_CHALLENGE_RESPONSE only if the submission dashboard requires an exact challenge body.

ChatGPT App

The hosted MCP server is also ready to connect as a ChatGPT App in developer mode:

Connector URL: https://mcp.future.video/mcp
App manifest:  https://mcp.future.video/chatgpt-app.json
Widget URI:    ui://future-video-studio/render-console-v1.html

The app exposes a small render console inside ChatGPT. It can prepare a render, submit account-wallet jobs, create no-account Link pay-per-render quotes, poll status, show payment links, and open the finished video when FVS returns a signed final_video_url.

ChatGPT connector setup:

  • Account mode: add X-FVS-Agent-Key as a secret connector header.
  • Pay-per-render mode: leave X-FVS-Agent-Key unset, or set X-FVS-Billing-Mode: pay-per-render.
  • Marketplace-linked mode: use the signed X-FVS-Marketplace-* headers from the account-linking gateway.

In ChatGPT, enable developer mode, create a connector, and use https://mcp.future.video/mcp as the connector URL. The tool metadata links the render tools to the FVS widget using the MCP Apps ui.resourceUri field and the ChatGPT openai/outputTemplate compatibility field.

A2A / Gemini Enterprise Preview

The hosted service also exposes a first A2A-compatible wrapper for the Google Cloud Marketplace / Gemini Enterprise Track 3 path:

https://mcp.future.video/.well-known/agent-card.json
https://mcp.future.video/message:send
https://mcp.future.video/tasks/{task_id}
https://mcp.future.video/tasks/{task_id}:cancel

The A2A layer uses the same X-FVS-Agent-Key account credential as the MCP server, marketplace-linked account headers, or paid quote mode. A message:send request can include either plain text brief parts or a structured fvs_render_request data part. Account and marketplace submissions create an FVS Agent API render and return an A2A task whose id is the FVS project_id. Paid quote submissions return HTTP 402 with an A2A task whose id is the quote_id; pay the returned Link/Stripe MPP payment_url, then poll /tasks/{quote_id}?claim_token=....

The Agent Card advertises the Official MCP Registry name, video.future/future-video-studio, so enterprise agents can see both the A2A delegation surface and the hosted MCP tool surface.

Billing Paths

  • Account mode: provide X-FVS-Agent-Key or FVS_AGENT_API_KEY. Renders charge the owning Future Video Studio account wallet under the same credit model and saved pipeline defaults as the web app.
  • Marketplace-linked mode: provide X-FVS-Marketplace-Account for a linked marketplace customer. Production deployments should also set X-FVS-Marketplace-Entitlement, X-FVS-Marketplace-Timestamp, and X-FVS-Marketplace-Signature from the account-linking gateway. The MCP server maps the marketplace account to a stored FVS Agent API key and charges the linked wallet.
  • Pay-per-render mode: omit account credentials or set X-FVS-Billing-Mode: pay-per-render. The server returns a Link/Stripe MPP payment quote, then exposes a claim-token status URL so the agent can retrieve the result.

Agents should get explicit user approval before submitting wallet-backed account renders.

Security Configuration

Marketplace account mapping is configured with environment variables on the MCP deployment:

$env:FVS_MARKETPLACE_MODE = "optional"
$env:FVS_MARKETPLACE_SHARED_SECRET = "replace-with-random-secret"
$env:FVS_MARKETPLACE_REQUIRE_SIGNATURE = "true"
$env:FVS_MARKETPLACE_ACCOUNT_KEYS_JSON = '{"acct_google_123":{"api_key":"fvs_live_replace_me","entitlement_id":"ent_123","plan":"starter","status":"active"}}'

The marketplace signature is HMAC-SHA256 over:

{unix_timestamp}.{marketplace_account_id}.{marketplace_entitlement_id}

Send it as X-FVS-Marketplace-Signature: sha256=<hex>. The default timestamp window is 300 seconds and can be adjusted with FVS_MARKETPLACE_SIGNATURE_MAX_SKEW_SECONDS.

For Cloud Run production deploys, pass FVS_MARKETPLACE_SHARED_SECRET_SECRET and FVS_MARKETPLACE_ACCOUNT_KEYS_JSON_SECRET to scripts/deploy_mcp_cloud.ps1 so the HMAC secret and account-key mapping come from Secret Manager instead of plain environment variables.

Tools

  • fvs_submit_render: submit a render request, optionally with public asset URLs
  • fvs_create_paid_render_quote: create a no-account Link payment quote
  • fvs_get_render_status: check a render by project_id or status_url
  • fvs_get_paid_render_status: check a paid render by status_url or quote_id plus claim_token
  • fvs_cancel_render: cancel a render by project_id or cancel_url
  • fvs_download_final_video: save a completed render from its signed URL
  • fvs_example_render_request: return an example scene render payload
  • fvs_open_chatgpt_app: open the ChatGPT app widget without creating a render

Downloading Final Renders

Use fvs_download_final_video only after fvs_get_render_status or fvs_get_paid_render_status returns a final_video_url for a completed render. If a signed URL has expired, call the relevant status tool again to refresh it.

The download tool:

  • performs an unauthenticated HTTPS GET to the signed final_video_url
  • writes the video bytes to output_path on the MCP server's local filesystem
  • creates missing parent directories
  • refuses to replace an existing file unless overwrite is true
  • does not require FVS_AGENT_API_KEY
  • does not spend wallet credits, create renders, cancel jobs, or modify remote state
  • may take minutes for large videos and uses a 600 second request timeout

Parameters:

  • final_video_url: an absolute HTTPS signed URL returned by a completed render status response. Do not pass arbitrary or untrusted URLs.
  • output_path: a local path visible to the MCP server process, such as C:/Users/me/Videos/fvs-result.mp4 or /tmp/fvs-result.mp4.
  • overwrite: defaults to false; set it to true only when replacing the destination file is intended.

Remote MCP Config

Account mode:

{
  "mcpServers": {
    "future-video-studio": {
      "url": "https://mcp.future.video/mcp",
      "headers": {
        "X-FVS-Agent-Key": "fvs_live_replace_me"
      }
    }
  }
}

Pay-per-render mode:

{
  "mcpServers": {
    "future-video-studio": {
      "url": "https://mcp.future.video/mcp"
    }
  }
}

Marketplace-linked mode:

{
  "mcpServers": {
    "future-video-studio": {
      "url": "https://mcp.future.video/mcp",
      "headers": {
        "X-FVS-Billing-Mode": "marketplace-linked-account",
        "X-FVS-Marketplace-Account": "acct_google_123",
        "X-FVS-Marketplace-Entitlement": "ent_123",
        "X-FVS-Marketplace-Timestamp": "<unix_timestamp>",
        "X-FVS-Marketplace-Signature": "sha256=<hmac>"
      }
    }
  }
}

Local Install

python -m pip install future-video-studio-mcp
python -m fvs_mcp_server

Example local stdio MCP config:

{
  "mcpServers": {
    "future-video-studio": {
      "command": "python",
      "args": ["-m", "fvs_mcp_server"],
      "env": {
        "FVS_AGENT_API_KEY": "fvs_live_replace_me",
        "FVS_AGENT_BASE_URL": "https://app.future.video"
      }
    }
  }
}

FVS_AGENT_API_KEY is optional. Leave it unset to use paid quote mode. FVS_AGENT_BASE_URL defaults to https://app.future.video.

Local Development

python -m pip install -e ".[dev]"
python -m pytest
python -m fvs_mcp_server

Run the HTTP transport locally:

$env:FVS_MCP_TRANSPORT = "streamable-http"
python -m fvs_mcp_server

Publishing

Build and check the package:

python -m build
python -m twine check dist/*

This package is intended to publish to PyPI as:

future-video-studio-mcp

The MCP Registry package verification marker is the hidden comment near the top of this README:

<!-- mcp-name: video.future/future-video-studio -->

Links

  • Future Video Studio: https://future.video
  • Agent API docs: https://future.video/api-docs
  • Hosted MCP manifest: https://mcp.future.video/server.json
  • Official MCP Registry name: video.future/future-video-studio

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured