What it does
Semantic Frame converts numerical data from NumPy arrays, Pandas and Polars objects, or Python lists into plain-language descriptions. Its analysis covers trend, volatility, anomalies, seasonality, distribution, step changes and data quality, computed with NumPy before being written as text. The publisher states a reduction of more than 95% in tokens compared with sending raw data points.
An LLM agent can be given a short summary in place of thousands of values, which the publisher presents as a way to avoid context overflow and to keep the model from estimating trends itself. The README also lists trading use cases, including drawdown, win rate, Sharpe ratio, market regime and strategy ranking, along with DevOps monitoring, sales analytics and IoT sensor data.
It is run as an MCP server and added to a client such as Claude Code or Claude Desktop through a configuration entry. The server exposes describe_data, describe_batch and describe_json, and the same functions can be called from Python through describe_series and describe_dataframe, which return text, JSON or a full result object.
Limitations and requirements
The README marks the describe_allocation trading tool with a warning symbol but does not explain it. The Advanced Tool Use features are labelled as beta. Setup requires editing the configuration file of the chosen MCP client.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.Anarkitty1/semantic-frame
Token-efficient semantic compression for numerical data. 95%+ token reduction.
v0.2.1As shown on registry.modelcontextprotocol.io when we last read it