
Semantic Frame
If you're building agents that analyze time series or tabular data, this server converts raw numerical arrays into token-efficient natural language summaries. It exposes three tools: describe_data for single series analysis, describe_batch for multiple columns, and describe_json for structured output. Instead of burning 2000 tokens on a thousand data points, you get a 50 word summary with trend classification, anomaly detection, volatility state, and seasonality metrics. The underlying analysis uses NumPy for deterministic stats, then translates results into LLM-friendly narratives. Works with Pandas, Polars, and NumPy inputs. Useful when you need Claude to reason about metrics, logs, or sensor data without hallucinating basic statistics or blowing your context window.
- GitHub stars
- 1
- tools
- 10
- transport
- STDIO
- last updated
- Dec 12, 2025
Tools
10 toolsPublic tool metadata for what this MCP can expose to an agent.
describe_dataAnalyze a single data series.
Analyze a single data series.
No parameter schema in public metadata yet.
describe_batchAnalyze multiple series at once.
Analyze multiple series at once.
No parameter schema in public metadata yet.
describe_jsonGet structured JSON output.
Get structured JSON output.
No parameter schema in public metadata yet.
describe_drawdownEquity curve drawdown analysis with severity.
Equity curve drawdown analysis with severity.
No parameter schema in public metadata yet.
describe_trading_performanceWin rate, Sharpe, profit factor metrics.
Win rate, Sharpe, profit factor metrics.
No parameter schema in public metadata yet.
describe_rankingsMulti-agent/strategy comparison.
Multi-agent/strategy comparison.
No parameter schema in public metadata yet.
describe_anomaliesEnhanced anomaly detection with PnL context.
Enhanced anomaly detection with PnL context.
No parameter schema in public metadata yet.
describe_windowsMulti-timeframe trend alignment.
Multi-timeframe trend alignment.
No parameter schema in public metadata yet.
describe_regimeMarket regime detection (bull/bear/sideways).
Market regime detection (bull/bear/sideways).
No parameter schema in public metadata yet.
describe_allocationPortfolio allocation suggestions.
Portfolio allocation suggestions.
No parameter schema in public metadata yet.
Sponsors
Readme
Semantic Frame
Token-efficient semantic compression for numerical data.
Semantic Frame converts raw numerical data (NumPy, Pandas, Polars) into natural language descriptions optimized for LLM consumption. Instead of sending thousands of data points to an AI agent, send a 50-word semantic summary.
The Problem
LLMs are terrible at arithmetic. When you send raw data like [100, 102, 99, 101, 500, 100, 98] to GPT-4 or Claude:
- Token waste: 1000 data points = ~2000 tokens
- Hallucination risk: LLMs guess trends instead of calculating them
- Context overflow: Large datasets fill the context window
The Solution
Semantic Frame provides deterministic analysis using NumPy, then translates results into token-efficient narratives:
Details
- Repository
- github.com/anarkitty1/semantic-frame
- Category
- Data & Analytics
- Package
semantic-frame- Transport
- STDIO
- Registry
- MCP registry, active