Overview

Work in progress. kestrel-chartkit is under active development and has not reached a stable release. Names, parameters, outputs and whole modules can still change fundamentally from one version to the next. Pin a release tag, and expect to adapt when you upgrade.

kestrel-chartkit is a Rust library for technical analysis. It streams price bars through indicators, classifies the market regime, combines indicator readings into one composite signal and returns plain data. Everything runs in-process on data you pass in; drawing is left to a renderer such as chartlet or kestrel-loom.

What it covers

• Indicators: 136 entries in the registry (catalog()), from moving averages and oscillators to volume profiles and market-structure detectors. All of them implement one streaming trait, and each is described by one definition: parameters with defaults, ranges and meaning (see the indicator reference).
• Market regime: classify_regime sorts the current market into bullish expansion, bearish expansion, consolidation or transition.
• Composite scoring: score_indicator turns one indicator reading into a subscore, aggregate_subscores weighs them into a direction, a permission grade and an ATR-based risk plan.
• Data for charts: a renderer-neutral scene model (viz::scene) for overlays, and every output as plain data for chartlet or kestrel-loom. See Visualising results.
• Structure and events: confirmed pivots and zones with a known event time, and detectors for failed breakouts, double bottoms and tops, bottom formation and divergences that alert on the bar the event becomes known.
• Analytics and evaluation: on-demand read-outs of the current market, forward paths of comparable situations, and what followed a signal: excursions, R-multiples, calibration, the deflated Sharpe ratio and a logistic model.

The crate also contains option and bond valuation, portfolio risk and stress scenarios. This documentation focuses on the areas above; the README and the API docs cover the rest.

Where it stops

• No data, no brokers. The crate is vendor-neutral: it does not fetch market data, connect to exchanges or ship market calendars. Where bars and instrument terms come from is your application’s business.
• No strategy. A composite signal is a rule-based summary of indicator readings. Its agreement is the share of subscores that agree, not a probability of success. Nothing in this crate or on this site is investment advice.
• Pre-1.0, in development. Root-level re-exports are the preferred API. Lower-level modules are public for advanced use. Nothing here is a stable foundation yet: expect breaking changes.
• Pure computation. The crate holds mathematics only: indicators, statistics, valuation formulas. It contains no credentials, no account or personal data and no connection to any service.

License

Business Source License 1.1 (BUSL-1.1). Free for non-commercial use, including production use in private, academic, non-profit and open-source projects that are not part of a commercial product or service. Commercial use requires a license from the Licensor. Each version converts to Apache-2.0 four years after publication. The LICENSE file is authoritative.

How the docs are organised

• Getting started: add the crate and stream a first indicator.
• Guides: indicators, market regime, composite scoring, visualising results, structure and events, analytics and evaluation.
• Reference: overview of the public entry points, the generated list of all indicators with their parameters, and where the API docs live.