Skip to content

MCP Overview

The Model Context Protocol (MCP) is an open standard for connecting LLMs to external tools and data sources. Thogits implements an MCP server that exposes its full feature set as callable tools, allowing AI assistants like Claude to create thogits, manage tags, search with filters, and more.

PropertyValue
Protocol version2024-11-05
Server namethogits-mcp
Server version0.1.0
Capabilitiestools

Thogits uses a dual-channel transport:

  • SSE (Server-Sent Events)GET /api/mcp/sse — server-to-client messages
  • JSON-RPC 2.0POST /api/mcp/message — client-to-server requests

All requests to /api/mcp/message require authentication via session cookies.

  1. Open SSE stream — Send GET /api/mcp/sse to establish a persistent connection.

  2. Receive endpoint event — The server sends an endpoint event with the POST URL:

    event: endpoint
    data: /api/mcp/message
  3. Send JSON-RPC requests — POST JSON-RPC messages to the endpoint URL. Responses arrive both as the HTTP response body and as SSE message events.

Add this to your .mcp.json to connect a client:

{
"mcpServers": {
"thogits": {
"type": "sse",
"url": "https://app.thogits.com/api/mcp/sse"
}
}
}

Every MCP session begins with an initialize request followed by an notifications/initialized notification.

{
"jsonrpc": "2.0",
"id": 1,
"method": "initialize",
"params": {
"protocolVersion": "2024-11-05",
"capabilities": {},
"clientInfo": {
"name": "my-client",
"version": "1.0.0"
}
}
}

Use tools/list to discover all available tools:

// Request
{
"jsonrpc": "2.0",
"id": 3,
"method": "tools/list",
"params": {}
}

The response contains an array of tool objects, each with name, description, and inputSchema.

All tool invocations use the tools/call method:

{
"jsonrpc": "2.0",
"id": 4,
"method": "tools/call",
"params": {
"name": "tool_name",
"arguments": { ... }
}
}

Successful responses return a content array with text results:

{
"jsonrpc": "2.0",
"id": 4,
"result": {
"content": [
{
"type": "text",
"text": "{ ... JSON result ... }"
}
]
}
}

Tool-level errors also return a content array but include "isError": true:

{
"jsonrpc": "2.0",
"id": 4,
"result": {
"content": [
{
"type": "text",
"text": "Error: Tag not found"
}
],
"isError": true
}
}

Protocol-level errors (unknown method, malformed request) use standard JSON-RPC error responses:

{
"jsonrpc": "2.0",
"id": 5,
"error": {
"code": -32601,
"message": "Method not found: unknown/method"
}
}

Thogits exposes 29 tools organized into these categories:

Tag Tools

add_tag get_tag edit_tag delete_tag get_tags

Create, read, update, delete, and search tag schemas.

Thogit Tools

add_thogit get_thogit edit_thogit delete_thogit get_thogits

Manage knowledge entries with full filter support.

Bulk Operations

bulk_add_thogits bulk_delete_thogits bulk_apply_tag bulk_remove_tag bulk_update_field_values

Batch create, delete, and modify thogits.

Relationship Tools

apply_tag_to_thogit update_tag_field_values remove_tag_from_thogit

Manage tag-thogit associations and field values.

View Tools

create_view get_view list_views update_view delete_view

Saved filter presets for organizing thogits.

Audit Tools

query_audit_log get_outdated_thogit_tags

Query event history and detect schema drift.

Lifecycle Tools

restore_thogit list_deleted_thogits purge_deleted_thogits

Restore soft-deleted thogits or permanently remove them.

Utility Tools

validate_filter explain_query render_react

Validate filters, explain queries, and render React components.

MethodDescription
initializeStart session, exchange capabilities
notifications/initializedConfirm initialization complete
tools/listList all available tools
tools/callInvoke a tool by name with arguments