TL-DPS

Provides a local MCP toolkit to parse Throne & Liberty combat logs into actionable DPS analytics via a local AI coach.

python

0

GitHub Stars

python

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "stalcup-dev-tl-dps-mcp": {
      "command": "python",
      "args": [
        "-m",
        "mcp_server"
      ],
      "env": {
        "PYTHONWARNINGS": "default"
      }
    }
  }
}

DPSCoach MCP Server lets you analyze Throne & Liberty combat logs by routing them through a local, read-only MCP server that returns actionable DPS metrics and explanations via a coach. It works offline, reads UTF-8 logs, and provides deterministic paths for most questions while keeping the UI responsive and safe.

How to use

You connect an MCP client to the server to analyze your TL combat logs. Start the MCP server, then configure your MCP client to point at that local process. You can ask the coach questions about your performance, request an analysis packet that includes run summaries and skill deltas, or run ad-hoc SQL-backed queries against your loaded log data.

Typical usage pattern:

Launch the MCP server from your CLI/MCP client: provide the same payload shape to all clients.
Point the client at the local server endpoint using the MCP JSON config snippet shown in the client setup example.
Send a request to analyze your logs (usually via the MCP tool’s run or analyze command) and receive a single analysis packet with summary metrics, top skills, and timeline deltas.
Optionally query the DuckDB event store using the provided MCP tools (for example, to fetch DPS metrics or event schemas) and view the results in your client.
If the LLM coach needs guidance, use deterministic handlers first; most questions will resolve to direct data without invoking the language model.

Available tools

analyze_dps_logs

Parses Throne & Liberty combat logs into structured event data for analysis.

query_dps

Executes read-only SQL-like queries against the in-memory DuckDB event store to retrieve DPS metrics, runs, and skill deltas.

get_events_schema

Returns the schema of the in-memory DuckDB events store used by the MCP server.

get_analysis_packet

Provides a single, comprehensive analysis packet containing run summaries, top skills, timeline buckets, skill deltas, and action levers.

coach_intent_router

Routes client questions to deterministic handlers or LLM-planned SQL queries based on detected intent such as RUNS, SKILLS, or CRIT_BUCKET_TREND.