The Vidmob MCP connector lets your team — or your own agents and applications — access Vidmob's creative data. Vidmob runs a standard remote MCP server, so it works with any MCP-compatible client: the major AI assistants and custom-built agents alike. In most cases all you need is our server URL.
This guide gives you the Vidmob-specific values and points you to each platform's own setup instructions, since those interfaces change frequently.
Before you start
You'll need admin permissions on your AI platform to add a custom connector / MCP server. (Most platforms don't let individual users add one themselves.)
After you set it up, each user signs in once the first time they use it — see the "After setup" section below.
What you'll need from Vidmob
Field | Value |
MCP Server URL | |
Authentication method | OAuth |
Header scheme (if asked) | Bearer |
Client ID + Client Secret | Whether you need these depends on your agent — some require them (e.g., Gemini Enterprise), others don't. Request from Vidmob if your platform asks for them (see below). |
Many platforms (such as Claude and ChatGPT) register automatically when you choose OAuth, so you won't need a Client ID or Secret. Others may ask for them — request those from Vidmob.
Already have an agent, or building your own? You can use OAuth as above, or authenticate with a Vidmob API key instead — often simpler for backend agents. See the "Custom or self-built agents" section below.
Setup by platform
The major hosted AI assistants are covered below as examples — the same approach works with any MCP-compatible client. Follow your platform's own documentation for the exact, current screens. Already have an agent, or building your own? See the "Custom or self-built agents" section below.
Claude (claude.ai / Claude Desktop)
Go to your connector/MCP settings and choose to add a custom connector.
Enter the Server URL: https://mcp.vidmob.com/mcp
Choose OAuth for authentication.
Save.
📄 Anthropic's instructions: https://support.claude.com/en/articles/11175166-get-started-with-custom-connectors-using-remote-mcp
ChatGPT
Create a new App / custom MCP connector.
Connection: choose Server URL and enter https://mcp.vidmob.com/mcp
Authentication: select OAuth (not "Access token / API key").
Header scheme: Bearer.
Create.
📄 OpenAI's instructions: https://help.openai.com/en/articles/11487775-apps-in-chatgpt
Google Gemini
Important: Custom MCP connectors are not supported in the consumer or workspace Gemini web app (gemini.google.com). To use the Vidmob MCP connector with Gemini, you must use one of the two developer-facing environments outlined below.
Option 1: Gemini Enterprise (Google Cloud)
Best for: Enterprise IT administrators managing team-wide AI agents.
A Google Cloud administrator must configure the connector as a custom data store inside the Google Cloud Console.
Request your credentials from Vidmob (see the instructions at the bottom of this page). You will receive a Client ID, Client Secret, Authorization URL, and Token URL.
Log into your Google Cloud Console and navigate to the Gemini Enterprise / Agent Builder platform.
Add the Vidmob MCP server as a custom connector using the base URL: https://mcp.vidmob.com/mcp along with the four credential values provided by Vidmob.
Finalize any organization-level IAM permissions or security policies required by your Google Cloud environment to authorize the external connector.
📄 Google's instructions: Set up a custom MCP server in Gemini Enterprise
Option 2: Gemini CLI (Local Terminal)
Best for: Individual developers, data scientists, and technical power users working locally.
Google's command-line AI tool natively supports connecting to external HTTP/SSE Model Context Protocol servers.
Ensure you have the Gemini CLI installed globally on your machine by running:
npm install -g @google/gemini-cliInitialize and authenticate your CLI by running gemini in your terminal and selecting your login method.
Add the Vidmob remote MCP server by executing the following command:
gemini mcp add --transport http vidmob [https://mcp.vidmob.com/mcp](https://mcp.vidmob.com/mcp)Trigger the OAuth authentication flow to securely link your Vidmob account data by running:
gemini mcp auth vidmob
📄 Google's instructions: Configuring MCP servers with Gemini CLI
How to Request Gemini Credentials
To get started with either option, please email support@vidmob.com. In your message, explicitly state whether you are setting up the connector for Gemini Enterprise (Google Cloud) or the Gemini CLI so our team can issue the correct metadata formats.
Custom or self-built agents
Because Vidmob runs a standard remote MCP server, it works with any MCP-compatible agent — including ones your team builds on frameworks like LangChain or Amazon Bedrock, or with your own code. Point your agent at the same server URL (https://mcp.vidmob.com/mcp) and choose whichever authentication fits your setup:
OAuth — use the standard OAuth flow described above, if your framework supports OAuth against an MCP server. Best when individual users sign in as themselves.
API key — for backend or server-to-server agents that you control, an API key is usually simpler: it's issued per organization, lives in your backend, and doesn't require an interactive login. Any Organizational Admin can generate a key directly in ACS at https://acs.vidmob.com/api-key-management. For how to generate and use it, see How do I set up the Creative Scoring and Creative Tags APIs (or the Vidmob API Overview for the bigger picture). Questions? Email support@vidmob.com.
After setup
Once the connector is added, each user connects it to their own account the first time they use it:
The user enables/connects the Vidmob connector in their AI assistant.
They're taken to the Vidmob login to sign in.
If they belong to more than one organization, they choose which one to use.
They approve access, and the Vidmob tools become available.
This is a one-time sign-in per user — they'll stay connected afterward.
Need help?
Contact support@vidmob.com — including to request Client ID / Secret credentials for any agent that requires them (such as Gemini Enterprise).