Your creative tools.
Inside your AI app.

Connect once. Create videos, images, and audio from your conversations.

Sign in with your Bandicam.ai account No API key required

Connect with Claude(Web, desktop & Claude Code)

  1. Copy the server URL

    Use these details when adding your connector.

    Connector nameBandicam AI
  2. Add a custom connector

    In the Claude app or on claude.ai, open Settings, then Connectors. Some versions list it under Customize.

    SettingsConnectorsAdd
    Add custom connectorConnector name: Bandicam AIMCP server URL: https://mcp.bandicam.ai/mcpEnter both, then click Continue.
  3. Sign in and enable

    Click Connect next to Bandicam AI if sign-in does not open on its own. Sign in to Bandicam.ai and allow access, then turn the connector on in a chat. Claude Code picks the connector up on its own — there is no command to run.

    Connect → sign in to Bandicam.aiAllow accessNew chat → + → Connectors → Bandicam AI on
Requirements & setup reference

Free Claude accounts can add one custom connector. Team and Enterprise may require an owner to add it first. Claude setup reference ↗

Claude Code loads your Claude connectors only when it is signed in with your Claude account, not with an API key. In the terminal, /mcp lists Bandicam AI once it is connected.

Prefer the command line? Run claude mcp add --transport http bandicam-ai https://mcp.bandicam.ai/mcp, then open /mcp › bandicam-ai and sign in in your browser. Claude Code MCP guide ↗

Menu labels may vary by client version and account.

Connect with Codex(Desktop app)

  1. Add an MCP connection

    In the Codex app, open Settings, select Plugins and then the MCPs tab. Open the Add menu, then choose Add MCP server.

    SettingsPluginsMCPsAddAdd MCP server
  2. Add the MCP server

    In Connect to a custom MCP, set Type to Streamable HTTP. Enter the name and server URL below, then click Save.

    NameBandicam AI
    TypeStreamable HTTP
  3. Sign in and connect

    Restart Codex if it asks. When Bandicam AI shows Authenticate, select it, sign in with your Bandicam.ai account and allow access.

    Start a Codex taskAsk to use Bandicam AIReview and approve actions
Requirements & setup reference

These steps are for the Codex desktop app. Add the remote server, then complete the Bandicam.ai browser sign-in. If you see Command to launch and Arguments, Type is still set to STDIO. Select Streamable HTTP to configure a remote server URL. Using the Codex CLI instead? Run codex mcp add bandicam-ai --url https://mcp.bandicam.ai/mcp, then codex mcp login bandicam-ai; the app, CLI and IDE extension share this setup. Menu labels may vary by app version. Codex MCP reference ↗

Connect with Cursor(Editor & Agent)

  1. Open MCP settings

    In Cursor, open Customize in the sidebar, select MCPs, then choose New MCP Server.

    CustomizeMCPsNew MCP Server
    Opens mcp.jsonKeep any servers already in the file.
  2. Add Bandicam AI

    Paste this entry into mcp.json and save.

    mcp.json
    {
      "mcpServers": {
        "bandicam-ai": {
          "url":
          "https://mcp.bandicam.ai/mcp"
        }
      }
    }
  3. Sign in once

    Click Needs login next to bandicam-ai and approve in your browser. Bandicam AI tools then show up in Agent chats.

    Needs login → sign in to Bandicam.aiAllow accessUse Cursor Agent
Requirements & setup reference

New MCP Server opens ~/.cursor/mcp.json, which applies to all projects. To use Bandicam AI in one project only, add the same entry to .cursor/mcp.json in that project. In older Cursor versions, find it under Settings › MCP & Integrations.

Cursor needs a remote MCP connection with OAuth support. Use the current app version and check your workspace settings if the server is unavailable.

Menu labels may vary by client version and account.

Connect with ChatGPT(Web)

  1. Open New Plugin

    In ChatGPT on the web, turn on Developer mode, then open Plugins and select the plus button next to the search box.

    SettingsSecurity and loginDeveloper mode
    Turn on Developer mode firstIt is near the bottom of the page.Then open Plugins → + → New Plugin.
  2. Add the MCP server

    In New Plugin, enter the name, then paste the server URL under Connection with Server URL selected. Keep Authentication set to OAuth.

    NameBandicam AI
    Authentication: OAuthIcon and Description are optional. Leave Advanced OAuth settings unchanged unless prompted.
  3. Create and authorize

    Read the custom MCP notice and check I understand and want to continue. Click Create, then complete any Bandicam.ai sign-in and access prompts.

    Sign in to Bandicam.ai and allow accessNew chat → + → More → Bandicam AISend the test prompt
Requirements & setup reference

Developer mode is available on the web for Plus, Pro, Business, Enterprise and Education accounts according to OpenAI's developer documentation. Workspace policies may restrict access or actions. Image and video generation require the Bandicam AI generation tools to be enabled and permitted.

If Create is disabled, check that Name and Server URL are filled in and the consent checkbox is selected. If sign-in does not finish, return to the plugin settings and reconnect.

Menu labels and rollout may vary. After you connect, Bandicam AI is also listed in Settings → Plugins, where you can manage or reconnect it. ChatGPT developer-mode reference ↗

Connect another MCP client(Remote MCP)

  1. Check client support

    Your client must support remote MCP Streamable HTTP and OAuth with dynamic client registration.

    Check its documentation before adding the server.

  2. Register the server

    Name the server Bandicam AI and enter the URL. Choose HTTP or Streamable HTTP for the transport.

    Connector nameBandicam AI
  3. Authorize your account

    Complete the browser login with your Bandicam.ai account. Review the requested permissions and allow access.

Requirements & setup reference

Clients without compatible OAuth support may not connect. Supported MCP revisions: 2025-06-18 and 2025-03-26.

Menu labels may vary by client version and account.

Check that you’re connected.

Ask your assistant to list your models and credit balance.

Use Bandicam AI to list the available models and show my credit balance. Do not generate anything yet.
Need help connecting?

Create with Codex

Start a task in the Codex app with Bandicam AI connected. Ask Codex to use Bandicam AI to create your references and video.

  1. Paste the example prompt into your Codex task. Ask for the model and credit estimate before approving generation.

  2. Attach both generated images, describe the motion, then review and approve the video estimate.

  3. View the result in your conversation or your Bandicam.ai library.

Codex MCP guide ↗
Example conversationCodex
You

Create two separate reference images with Bandicam AI: a green explorer robot on a neutral background, and a pink alien landscape with a ringed planet. Show the credit estimate first and wait for approval.

Codex

Review the suggested model, settings and credit estimate.

You

Go ahead with those settings.

Generated reference 1: mint green explorer robot on a dark studio backgroundReference 1 · Character
Generated reference 2: pink alien dunes beneath a ringed planetReference 2 · Environment
You
Reference 1 Reference 2

Use both reference images to make a 5-second video: the robot walks through the alien landscape and waves. Keep its design and the landscape consistent. Show the credit estimate first.

Codex

Use Reference 1 for the robot and Reference 2 for the environment. Review the video model, settings and credits.

You

Approved. Bring the robot to life.

One workflow. More possibilities. Try videos, voiceovers, music and sound effects with the same steps.

Ask it. Create it.

Video, images and audio. All from your conversation.

Product videos

Turn a product idea into a cinematic video.

Original images

Create characters and worlds in your own style.

Voice & sound

Add narration, music and sound to your story.

Create videos

Turn a prompt or image into a video.

Generate images

Bring your ideas to life in any style.

Add voice & sound

Create narration, music and sound effects.

Edit & transform

Upscale, swap faces or remove backgrounds.

Check the cost

See the credit estimate before you create.

Reuse your creations

Find past results in your Bandicam library.

Explore all AI tools →

MCP connection FAQ

Do I need an API key?

No. Add the server URL in a supported client, then sign in to your Bandicam.ai account using Sign in to Bandicam.ai and allow the requested access. OAuth handles the connection.

How much does it cost?

Generation consumes the same Bandicam.ai account credits as the web app. The amount depends on the model and settings. Ask your assistant to call quote_job before submitting a job. Your AI app's own subscription requirements are separate.

Can I try it with a free Bandicam.ai account?

Yes. Free Bandicam.ai accounts are eligible, and new accounts can use their welcome credits for supported generations. Your chosen AI client may have its own plan or workspace restrictions.

Where are my creations saved?

Your generated media is saved to your account library. Sign in with the same Bandicam.ai account on bandicam.ai to find it. Your assistant can also look up previous creations for reuse.

How do I disconnect an assistant?

Sign in to bandicam.ai, open the account menu, and go to Connected apps. Revoke the connection you want to remove. You can also remove the server from your client's settings.

What should I check if I can't connect?

The server won't connect: Check that the server URL is https://mcp.bandicam.ai/mcp, including /mcp. Your client needs remote MCP support with Streamable HTTP and OAuth.

Sign-in didn't finish: Return to the connector settings and try connecting again. Complete the Bandicam.ai sign-in and allow the requested access in your browser.

Connected, but no tools appear: Start a new conversation and enable Bandicam AI in your assistant's tools or connector menu. Then use the test prompt above.

The connector option is missing: Check the Requirements & setup reference for your selected client and any workspace restrictions.

If it still fails, contact us with the client name and the error message you see.

Choose your assistant and get set up.

Copy the server details and follow the steps for your preferred client.