Skip to main content

Connecting Streamtime to Your AI Assistant (via MCP)

Connect Streamtime to Claude, ChatGPT, Microsoft Copilot, or your own custom connection through our MCP Server.

🕰️ Est. Time: 5–15 minutes, depending on your assistant
🎯 Goal: Get your AI assistant talking to your Streamtime account
🔑 Requirements: An active Streamtime login, and an AI assistant that supports remote MCP servers. Plan requirements vary by platform, so check each section.

ℹ️: These connection steps were last checked in August 2026. AI platforms change their setup flows often, so if something doesn’t match what you see, check your platform’s own help docs for the latest steps.

⚠️: When you connect an AI assistant, you’re asking us to share your Streamtime data with the provider of that assistant. You choose the provider. They aren’t one of our service providers, and we have no agreement with them covering your data, so what they do with it is governed by your agreement with them. Read their terms first, and bear in mind your account holds information about your team and your clients’ people.

AI assistants can also misread requests and produce inaccurate or incomplete output. Streamtime isn’t responsible for the accuracy of AI-generated output, so review anything that matters before you act on it, especially financial figures, edits and deletions.

Once connected, your assistant can read from and write to your Streamtime account on your behalf. If you haven’t already, read what the connector can and can’t do before setting it up.

🔑 What you'll need

Whichever assistant you use, you’ll need the following.

The server address: https://mcp.streamtime.net

How sign-in works: The connector uses OAuth 2.0 with Dynamic Client Registration (DCR). In plain terms, most assistants only need the server address above. Paste it in, a Streamtime login window opens, you sign in, and you’re connected.

You don’t need to generate an API key, client ID or client secret. If an assistant asks for them, you can usually leave those fields blank.

The connector signs in as you and inherits your Streamtime permissions exactly. It can’t see or do anything you can’t already see or do in the app. Read more about permission levels here.

🔒 How is my data kept secure?

Data between Streamtime and your AI assistant is sent over an encrypted connection (SSL), in the same way as our API.

Once your data reaches your AI assistant, what happens to it is up to the provider of that assistant. That includes whether they use it to train their models, which often depends on your plan with them. Free plans are generally more likely to use your data for training than paid plans like Pro or Team. Providers update these policies often, so check your plan’s current terms and privacy settings before you connect.

Streamtime does not use customer data to train any AI models.

🤖 Connecting to Claude

Custom connectors are available on all Claude plans: Free, Pro, Max, Team and Enterprise. Free plans can add one custom connector. Claude’s custom connectors are still in beta on their side.

On Free, Pro or Max:

  1. Go to Customize > Connectors (or open claude.ai/customize/connectors).

  2. Click Add custom connector.

  3. Paste https://mcp.streamtime.net as the remote MCP server URL.

  4. Click Add, then click Connect.

  5. A Streamtime sign-in window opens. Log in to authorise the connection. Leave the optional OAuth Client ID and Client Secret fields (under Advanced settings) blank. If you have more than one Streamtime account, check you’ve selected the one you want to connect.

  6. You’re connected. To test it, ask Claude to run the Streamtime health check in any chat.

On Team or Enterprise: an Owner or Primary Owner adds the connector for your organisation first:

  1. Go to Organization settings > Connectors.

  2. Click Add, hover over Custom, then select Web.

  3. Paste https://mcp.streamtime.net and finish adding the connector.

After that, each team member finds Streamtime under Customize > Connectors (it has a Custom label), clicks Connect and signs in with their own Streamtime login.

💬 Connecting to ChatGPT

ChatGPT connects to custom MCP servers through Developer mode on the web app.

It’s available on Plus and Pro personal accounts. On Business, Enterprise and Edu, a workspace admin has to turn it on first (and can restrict it), under Workspace Settings > Permissions & Roles > Connected Data. Free accounts can’t add custom MCP servers.

  1. Turn on Developer mode. On personal accounts, this currently lives under Settings > Security and login > Developer mode.

  2. Head to Plugins and add a new plugin. Paste https://mcp.streamtime.net and set authentication to OAuth.

  3. A Streamtime sign-in window opens. Log in to authorise the connection. Leave the optional OAuth Client ID and Client Secret fields (under Advanced settings) blank. If you have more than one Streamtime account, check you’ve selected the one you want to connect.

  4. Turn the connector on in your conversation. ChatGPT may ask you to confirm each action, so approve them as they come up.

🛠️: ChatGPT has historically leaned on search and fetch style tools, and can be less consistent than other assistants at using a server’s full set of tools. If it isn’t using the Streamtime action you expect, ask for it by name. You can also check the connection by asking ChatGPT to run the Streamtime health check.

💡: Google Gemini and Microsoft Copilot can both connect, but neither is as quick to set up as Claude or ChatGPT. Both take more work, and both put custom MCP access behind specific plan tiers or admin involvement. That’s true for any MCP server, not just Streamtime, so the effort multiplies if you’re connecting other tools too.

For the fastest way to try the connector, use Claude or ChatGPT. If your organisation is standardised on Google or Microsoft, the next two sections will get you started.

🔷 Connecting to Google Gemini

Custom MCP connectors in Gemini are only available on Enterprise plans. Lower tiers can’t add a custom remote MCP server, and in most cases an administrator needs to turn it on for your organisation first.

  1. Confirm you’re on an Enterprise plan, with custom connectors enabled by your administrator.

  2. Add a custom MCP connector and enter https://mcp.streamtime.net as the server URL.

  3. When prompted, authenticate with OAuth 2.0. A Streamtime sign-in window opens. Log in to authorise the connection, and check you’ve selected the right account.

For current setup steps and admin requirements, see Google’s Gemini for business documentation. As with every other assistant, the connection acts as the signed-in Streamtime user and is bound by that user’s permissions.

🟦 Connecting to Microsoft Copilot

“Copilot” covers a few different products, and its MCP support is aimed at people building agents or administering a tenant, not at the everyday Copilot chat box.

You’ll need a Copilot Studio account to build the agent, then publish it to your Copilot through Microsoft. Microsoft has more on agents and Copilot Studio here.

Build the agent in Copilot Studio

  1. Open your agent and go to Tools > Add tool > New tool > MCP.

  2. Enter https://mcp.streamtime.net as the server URL.

  3. The MCP configuration wizard opens. Name and describe the connection however you like, then select OAuth 2.0 with Dynamic Client Registration.

  4. You may need to finish the connection by clicking the sign-in button and logging in to Streamtime. As with other tools, check you’ve selected the right account.

Copilot connects to MCP servers over the Streamable HTTP transport, and most of this needs Power Platform or admin permissions. You may need to ask your Microsoft administrator for the right permissions and environment setup.

Publish the agent to Microsoft 365 Copilot

Building the agent in Copilot Studio is only half the job. You then need to publish it so it shows up in Microsoft 365 Copilot. For the steps, see Microsoft’s guide on publishing a Copilot Studio agent to Microsoft 365 or Teams.

🔧 Connecting a custom or other client

Any MCP-compatible client can connect the same way, whether that’s a coding tool like Cursor or VS Code, or your own application.

Point the client at:

  • Server URL: https://mcp.streamtime.net

  • Transport: Streamable HTTP (SSE is deprecated and not supported)

  • Authentication: OAuth 2.0 with Dynamic Client Registration. No pre-issued credentials are needed. The client registers itself and sends you through a Streamtime sign-in.

Beyond that, follow your client’s own instructions for adding a remote MCP server. As with every other assistant, the connection acts as the signed-in Streamtime user and is bound by that user’s permissions.

ℹ️: 🤝 Official partner listings are on the way. We’ve applied to be listed as an official partner in the directories of some major AI platforms. Once approved, you’ll be able to add Streamtime straight from the platform’s built-in directory instead of pasting the server address. This guide covers the manual setup that works today.

Related Articles

Related Articles
Did this answer your question?