Awesome MCP › Other MCP Servers

optuna/optuna-mcp

⭐ 84 Python repository created 2025-06-10

Optuna MCP Server is maintained by the Optuna project and exposes its hyperparameter optimisation framework through MCP. Tools create and manage studies, suggest parameter values for the next trial, report the resulting objective value, and query the state of an optimisation — best trial, parameter importances and the history — so a model can drive an optimisation loop step by step. It also exposes Optuna's visualisation and analysis functions, which supports the second documented use case: interactive examination of a completed study through chat instead of a notebook. The third use case is the most interesting and is what distinguishes this from a plain library wrapper — optimising the inputs and outputs of other MCP tools. Because the assistant can call another server's tool, observe a result and feed it back as a trial outcome, Optuna can tune parameters of anything reachable over MCP, not only a machine-learning model in the same process. It is Python, requires 3.12 or 3.13, is published on PyPI as optuna-mcp, MIT licensed, with tests in continuous integration. The audience is machine-learning practitioners already using Optuna who want tuning and result analysis in conversation, and people experimenting with automated tuning of agent pipelines.

https://github.com/optuna/optuna-mcp

mcpoptunahyperparameter-optimizationpythonmachine-learningexperiment-tracking