ATars MCP

What aTars MCP MCP server does

aTars MCP MCP server gives AI agents access to a locally maintained cryptocurrency analytics dataset. The dataset contains up to 90 days of one-minute OHLCV candles and more than 40 derived technical indicators. Supported analysis covers price, momentum, trend, volatility, and volume.

The server also provides a separate daily sentiment dataset sourced from Perplexity AI. Sentiment tools return Bullish, Bearish, or Neutral classifications, numeric scores, historical values, rolling statistics, streaks, and composite verdicts.

How it works

Start with get_available_symbols when the supported market assets are unknown. get_data_info can then report a token's date range, candle count, freshness, and available feature names. These discovery steps are useful because the server's local dataset is authoritative for the symbols and features it exposes.

For current technical data, get_latest_features returns the latest values across the available indicators. Historical requests can use get_features_range, get_indicator, or get_multi_indicator, with selectable resolutions such as one minute, one hour, four hours, and one day. Category-specific requests are handled by get_category_features and get_category_summary.

Technical signal tools classify the latest data as Bullish, Bearish, or Neutral using indicators including RSI, MACD, EMA crossovers, ADX, Stochastic, and volume signals. get_signal_history provides hourly signal results over a historical window. These are indicator-based signals, not trading guarantees.

Sentiment analysis follows a different data path. Use get_sentiment_symbols to find tokens with stored sentiment data, then choose current readings, history, rolling trends, or composite sentiment with the corresponding sentiment tools.

Tools and capabilities

• Inspect available crypto symbols and dataset metadata.
• Retrieve current, historical, grouped, or side-by-side technical indicators.
• Review individual indicator histories for up to 90 days.
• Generate technical signal summaries and hourly signal histories.
• Calculate category statistics such as mean, minimum, maximum, standard deviation, latest value, and direction versus average.
• Profile indicators with null rates and Pearson correlation with close price.
• Export technical data as CSV or JSON, or request an ML-ready feature matrix.
• Retrieve daily Perplexity-sourced sentiment, sentiment history, momentum, streaks, and composite verdicts.

Limitations and notes

Historical technical queries return at most 500 rows per symbol through get_features_range. Several tools impose lookback limits: indicator and feature requests generally allow up to 90 days, while category features allow up to 30 days and signal history up to 30 days. Resampling choices vary by tool.

The technical sentiment-style signal is computed from local indicators, while the separate news sentiment dataset is sourced from Perplexity AI. Neither should be treated as a substitute for independent research or risk management. The supplied material does not specify authentication, environment variables, licensing, or client-specific setup details for the remote MCP endpoint.