Connecting to our Chargebee MCP server

The Model Context Protocol (MCP) allows you easily query your Chargebee data from within your favourite AI chatbot by simply connecting to our MCP server. > **What is MCP?** MCP is a globally accepted protocol for connecting AI clients to external tools where they obtain additional context for their conversations. In this case, that additional context is the data from your cloud services. MCP uses the latest standard security protocols such as OAuth2 and PKCE, so your connection is extremely secure. ### How to connect to your Chargebee MCP Server First, you'll need the address of our MCP server: **https://api.synchub.io/mcp** This is usually the only thing you'll need in order to connect. Some MCP Clients such as Copilot Studio require that you configure authorisation options is well. For these cases the SyncHub MCP server uses OAuth2.0 and supports Dynamic Discovery. The most recent MCP specification version we support is "2025-06-18". The exact instructions for adding these values depend on your AI tool, so please refer to their up-to-date documentation. However, we also have dedicated walkthroughs for [Claude](/connectors/chargebee/knowledge-base/claude) and [ChatGPT](/connectors/chargebee/knowledge-base/chatgpt). ### What does the our MCP connector offer? Our MCP connector provides access to your data in three ways: 1. Query your data using your natural language, for example _"Have we forgotten to issue any invoices?"_ or _"Help me forecast this year's sales"_ or _"Let me know if Jim is late to work next week"_ 2. For repeatable & predictable results, use our [Query Editor](/connectors/chargebee/knowledge-base/snippets) to build a SQL script and have your AI execute it on demand 3. Building on from the above, if your query is intensive or complex, our MCP server has access to your [Insights](/connectors/chargebee/knowledge-base/Insights) results, too #### Restricting access to MCP functionality The above features can all be enabled/disabled individually when you connect your MCP client to our server. For example, you may create a connection that allows users to query all the data via natural language (1), or you may want much tighter control and only release subsets of data via the Queries (2) and Insights (3). ### Restricting or segmenting data to your users There are two main ways in which you may wish to segment your data - within a single instance of Chargebee, or across multiple Chargebee connectors. #### Restricting access within a single Chargebee dataset Sometimes you may have sensitive data in parts of your Chargebee dataset, which you don't want your MCP clients to see. In these cases, you have a few options: - Remove option (1) above (the natural language feature) and provide access only via the queries (2) and insights (3) that you control - Our [obfuscation module](/connectors/chargebee/knowledge-base/obfuscation) allows you to prevent sensitive data from entering your dataset in the first place (note: this applies at the dataset level, not at the MCP level - so _nobody_ will be able to see this data, even in your reports, as it is literally never entered into your dataset) > For avoidance of doubt, these restrictions are applied at the _MCP Connection Level_ and are _not_ dependent on the user (i.e. a human or agentic workflow) which is accessing the MCP connection. See the section below for an explanation of our account-level security model. #### Restricting access across multiple Chargebee datasets Regardless of which features you activate for your specific MCP connection, please note that _all_ of your MCP clients (i.e. humans or agentic workflows) will have access to _all_ the features. This is because we authenticate your MCP connection at the _account_ level, rather than the individual _user_ level. This means that the only way you can separate access between the two Chargebee connectors is to create a **second SyncHub account**, loading a different connection into each (this security model maps with our overall connector-access model within SyncHub). > Creating a second account is not as burdensome or costly as it may first appear, because our pricing model lets you configure and pay for each connection separately anyway, and our app makes it very easy to switch between accounts within the same login. ### Building a dashboard Check out [this blog article](https://www.synchub.io/blog/aicorner-using-ai-to-build-business-dashboards) to see how easy it is to build dashboards using AI. > **Recommended:** for security peace of mind, and to remove any restrictions around our AI use, don't forget to also [connect your own AI model](/connectors/chargebee/knowledge-base/byoai).