home / mcp / kibela mcp server
Kibela's official MCP Server
Configuration
View docs{
"mcpServers": {
"kibela-kibela-mcp-server": {
"command": "docker",
"args": [
"run",
"-i",
"-e",
"KIBELA_ORIGIN",
"-e",
"KIBELA_ACCESS_TOKEN",
"ghcr.io/kibela/kibela-mcp-server"
],
"env": {
"KIBELA_ORIGIN": "https://your-subdomain.kibe.la",
"KIBELA_ACCESS_TOKEN": "***"
}
}
}
}Kibela MCP Server lets you access Kibela data through the Model Context Protocol (MCP) from your preferred MCP client. It runs locally and exposes tools to work with Kibela notes, folders, comments, and more, so you can query, organize, and create content with AI assistance directly from your MCP client.
You connect an MCP client like Claude Desktop or VSCode to the Kibela MCP Server running on your machine. The server exposes a set of public MCP tools for querying notes, folders, comments, and related data. Use these tools to search Kibela notes, retrieve notes by path or URL, create or update notes, manage folders, and add comments. You can also run prompts to review or explore notes. Configure your client to point at the Kibela origin and supply your Kibela access token to authorize requests.
Prerequisites you need before starting: Docker is installed and running. You must have a Kibela access token issued. You also need an MCP client installed to interact with the server.
Choose one of the following runtime methods to start the Kibela MCP Server.
{
"mcpServers": {
"kibela": {
"command": "docker",
"args": [
"run",
"-i",
"-e",
"KIBELA_ORIGIN",
"-e",
"KIBELA_ACCESS_TOKEN",
"ghcr.io/kibela/kibela-mcp-server"
],
"env": {
"KIBELA_ORIGIN": "https://your-subdomain.kibe.la",
"KIBELA_ACCESS_TOKEN": "***"
}
}
}
}If you prefer not to use Docker, set the script path as the execution command and ensure the path is absolute.
{
"mcpServers": {
"kibela": {
"command": "/path/to/kibela-mcp-server/bin/cli.mjs",
"env": {
"KIBELA_ORIGIN": "https://your-subdomain.kibe.la",
"KIBELA_ACCESS_TOKEN": "***"
}
}
}
}You can customize tool descriptions and prompts by providing a JSON file and mounting it into the container. Create a file with your new tool and prompt descriptions, then mount it at the expected path inside the container so the server can read the customization at startup.
{
"tools": {
"search_kibela_note": {
"description": "New description"
}
},
"prompts": {
"review": {
"prompt": "New review prompt"
}
}
}To apply the customization with Docker, mount the configuration file into the container at the following path and ensure the server starts with access to it.
{
"mcpServers": {
"kibela": {
"command": "docker",
"args": [
"run",
"-i",
"-e",
"KIBELA_ORIGIN",
"-e",
"KIBELA_ACCESS_TOKEN",
"-v",
"/path/to/kibela-mcp-server-config.json:/usr/src/app/kibela-mcp-server-config.json",
"ghcr.io/kibela/kibela-mcp-server"
],
"env": {
"KIBELA_ORIGIN": "https://your-subdomain.kibe.la",
"KIBELA_ACCESS_TOKEN": "***"
}
}
}
}Ensure Docker is running before starting the server. If you encounter authentication issues, double-check that your Kibela origin URL and access token are correct and that the token has the necessary scopes.
Search notes with multiple filters such as keywords, resource types, co-editing, update dates, group and folder filters, liker IDs, and sort order.
Retrieve a note by its Relay ID.
Fetch a note using its path or URL.
Get notes within a specified folder with optional pagination.
Create a new note with a title, content, and metadata like groups and folders.
Update the content of an existing note.
Search folders by keyword.
Retrieve a folder by its Relay ID.
Fetch a folder by path or URL.
List folders with optional pagination.
Create a new folder within a group.
Move a note from one folder to another.
Attach a note to a folder.
Create a comment on a note.
Create a reply to a comment.
Fetch Kibela groups.
Get feed sections for a specific kind and group.