Flowcore Platform MCP Server
About
MCP server for managing and interacting with Flowcore Platform
Details
- Author
- flowcore-io
- GitHub stars
- 9
- Downloads
- 164
- Categories
- Other
Jump to
- Standardized MCP interface for Flowcore Platform resources
- Run without installation using npx
- Supports both command-line arguments and environment variables
- Built with Bun for fast JavaScript runtime
- Open source with community support via Discord
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
Flowcore Platform MCP ServerCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
Run directly using npx without installation: npx @flowcore/platform-mcp-server --username <username> --pat <pat>. Alternatively, install globally with npm install -g @flowcore/platform-mcp-server and then run platform-mcp-server --username <username> --pat <pat>. Replace <username> and <pat> with your Flowcore username and Personal Access Token. Environment variables USERNAME and PAT can also be used instead of command-line arguments.
get_tenant
Get a tenant
get_data_core
Get a data core
get_flow_type
Get a flow type
get_event_type
Get an event type
list_tenants
List all tenants I have access to
list_data_cores
List all data cores for a tenant
list_flow_types
List all flow types for a data core
list_event_types
List all event types for a flow type
get_event_type_info
Get event information about an event type, like first and last time bucket and 5 example events. Sensitive data should not be needed as it is masked in a correct format in the response, but if you need it ask the user if they are sure they want to include sensitive data, is **false** by default. **DO NOT** fetch sensitive data from the event type on your own initiative, if you deem it necessary to fetch sensitive data, **ALWAYS** ask the user if they are sure they want to include sensitive data, is **false** by default
get_events
Get events for an event type, this can be paginated by using the cursor returned from the previous call. This is good for getting the payload of the events to inspect them. You can pageinate by using the cursor returned from the previous call. You can also filter by event id, time bucket, from event id, to event id, order, and page size. The order is ascending by default, and the page size is 500 by default. The cursor returned can be called nextCursor, it is a stringified object or a string, you need to pass it as is to paginate. Sensitive data should not be needed as it is masked in a correct format in the response, but if you need it ask the user if they are sure they want to include sensitive data, is **false** by default. **DO NOT** fetch sensitive data from the event type on your own initiative, if you deem it necessary to fetch sensitive data, **ALWAYS** ask the user if they are sure they want to include sensitive data, is **false** by default
get_time_buckets
Get time buckets for an event type, this is useful for getting the time buckets for an event type, and then using the get_events tool to get the events for a specific time bucket. The time bucket is in the format of YYYYMMDDhhiiss, normally the ii and ss are 0000. It can be paginated by using the cursor returned from the previous call.
create_data_core
Create a data core in a tenant
update_data_core
Update a data core
update_flow_type
Update a flow type
update_event_type
Update an event type
create_flow_type
Create a flow type in a data core
create_event_type
Create an event type in a flow type
request_delete_event_type
Request to delete an event type, this will delete all events in the event type, this is irreversible. If wait for delete is true the tool will wait up to 25 seconds for the event type to be deleted, if not it will return immediately, you have to use the get_event_type tool to check if the event type is deleted.
request_truncate_event_type
Request to truncate an event type, this will delete all events in the event type, this is irreversible. If wait for truncate is true the tool will wait up to 25 seconds for the event type to be truncated, if not it will return immediately, you have to use the get_event_type tool to check if the event type is truncated.
request_delete_flow_type
Request to delete a flow type, this will delete all events in the flow type, this is irreversible. If wait for delete is true the tool will wait up to 25 seconds for the flow type to be deleted, if not it will return immediately, you have to use the get_flow_type tool to check if the flow type is deleted.
request_delete_data_core
Request to delete a data core, this will delete all events in the data core, this is irreversible. If wait for delete is true the tool will wait up to 25 seconds for the data core to be deleted, if not it will return immediately, you have to use the get_data_core tool to check if the data core is deleted.
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"flowcore platform mcp server": {
"mcp-flowcore-platform": {
"command": "npx",
"args": [
"@flowcore/platform-mcp-server",
"--username",
"<username>",
"--pat",
"<pat>"
]
}
}
}
}McpServers
{
"mcp-flowcore-platform": {
"command": "npx",
"args": [
"@flowcore/platform-mcp-server",
"--username",
"<username>",
"--pat",
"<pat>"
]
}
}Flowcore Platform MCP Server
A Model Context Protocol (MCP) server for managing and interacting with the Flowcore Platform.
If you're curious about how it works, you can check out our video, where we set it up and demonstrate what it can do.
> If you like this solution, but would like to reduce the hallucinations, reduce token usage, increase the amount of data you can look through and also speed up querying exponentially - then we recommend you also check out our local read model mcp server,
Usage with npx
You can run this package directly using npx without installing it:
npx @flowcore/platform-mcp-server --username <username> --pat <pat>
Replace <username> and <pat> with your Flowcore username and PAT (Personal Access Token).
Installation
If you prefer to install the package globally:
npm install -g @flowcore/platform-mcp-server
Then run it:
platform-mcp-server --username <username> --pat <pat>
Development
To install dependencies:
bun install
Run the project directly with Bun:
bun run src/index.ts --username <username> --pat <pat>
Building
Build the project:
bun run build
Run the built project:
node dist/cli.js --username <username> --pat <pat>
Environment Variables
| Variable | Type | Description | Default | Required |
|----------|------|-------------|---------|----------|
| USERNAME | string | Flowcore username | - | ✓ |
| PAT | string | Flowcore PAT (Personal Access Token) | - | ✓ |
About
This project uses the Model Context Protocol (MCP) to provide a standardized interface for interacting with the Flowcore Platform. It allows AI assistants to query and manage Flowcore resources through a structured API.
Originally created using bun init in bun v1.2.3. Bun is a fast all-in-one JavaScript runtime.
Questions
if you have any questions or cool ideas, then feel free to join our Discord community. Otherwise you can find us on all major social platformsSign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
