StockSharp: a wide C# trading stack whose connector code and releases both live elsewhere

The newest tagged release is from March 2020 while master keeps moving

The GitHub release list is not a complete inventory of StockSharp distribution. StockSharp publishes versioned NuGet packages, including components updated in 2026. Check the package and version relevant to your integration, its release notes and compatibility, or pin a specific source commit if building from the repository. Do not infer package age from the newest GitHub release alone.

Connector source code is not in this repository

The connection layer lives somewhere else. Connector source code and the full connector list are both stated to be available in the separate StockSharp Connectors repository, and there is no connector directory in this tree, although the root does hold Fix.Core and Fix.Dialects for the FIX side of the protocol work.

The named connections span crypto venues such as Binance, Kraken, Coinbase, OKX, KuCoin, Deribit, HTX and Bitfinex, the MT4 and MT5 terminals, FIX and FAST, PolygonIO, Trading Technologies, Alpaca, BarChart, CQG, E*Trade, IQFeed, Interactive Brokers, LMAX, Oanda, FXCM, Rithmic, cTrader, DXtrade and QuantFEED. The consequence for a reader is direct: you cannot audit or fork a connector from this clone, and whether a given venue supports the market data types your strategy subscribes to is answerable only from the per-exchange documentation, which is a separate page per venue.

The sample strategy ignores every candle that is still forming

The connector example creates a connector, looks up a security by id, and subscribes to a one minute timeframe:

var connector = new Connector();
var security = connector.LookupById("AAPL@NASDAQ");

The handler then filters on candle state, and that filter is the mechanism worth reading rather than the strategy logic. It returns early unless the subscription matches and the candle state is CandleStates.Finished, so nothing happens while a bar is still forming, and it decides direction by comparing the close price to the open price before registering an order of type `OrderTypes.Market`. A one-minute subscription filters for finished candles, but the example does not establish deduplication or guarantee an execution price. Market orders can experience slippage; validate event handling and risk controls before any live use. The API is pitched at high frequency strategies with direct exchange access through DMA.

Four solution files decide which StockSharp you build

The root carries StockSharp.slnx alongside StockSharp_CI.slnx and StockSharp_Tests.slnx, and the samples area has its own Samples.sln and Samples.slnx. There is no single obvious entry point, and the choice changes what gets compiled.

Under Samples, numbered folders run from 01_Basic through 10_CrossPlatform, with a separate Common_Avalonia directory beside props files that split the sample set five ways: common_connectors.props, common_samples.props, common_samples_avalonia.props, common_samples_net.props and common_samples_netwindows.props. A netwindows group sitting next to a net group and an Avalonia group means the samples are not uniform across operating systems. The consequence for someone on Linux or macOS is that whether a given sample compiles depends on which props file it references, so the usual first step is to read the imports and see.

Analytics appears in three languages beside a C# core

The tree holds Algo.Analytics, Algo.Analytics.CSharp, Algo.Analytics.FSharp and Algo.Analytics.Python as four separate directories, while the rest of the stack stays C#: Algo.Indicators, Algo.Strategies, Algo.Testing, Algo.Import, Algo.Export, Algo.Compilation and Algo.Gpu. Sitting alongside them are MatchingEngine, Messages, BusinessEntities, Charting.Interfaces, Reporting, Diagram.Core, and a Localization trio made of Localization, Localization.Generator and Localization.Langs.

The presence of Python and F# analytics projects inside a C# platform says the analysis side is not locked to one language, and the directories hint at a matching engine, a GPU path, and generator projects for media and localization. What none of it tells you is which of these components are complete, which are wired into the sample solutions, or whether the Python analytics project is a supported dependency or an experiment. A reader who assumes a uniform supported stack from the folder names will be disappointed, so check each component against the samples before designing around it.

License metadata reads NOASSERTION beside a LICENSE file

The repository's license field reports NOASSERTION, the value recorded when a classifier cannot decide the terms, while the root of the tree contains both a LICENSE file and a NOTICE file. The licensing information exists on disk but is not machine-readable from the repository header.

The consequence shows up in tooling. Any dependency scanner, license allow list, or automated compliance report pointed at this project treats it as unknown rather than reading what LICENSE says, and the NOTICE file is not consulted automatically either. For a team deciding whether the C# API can ship inside a commercial product, the terms have to be read from those two files by a person. Given how many project directories sit in the tree, a scanner that skips it on metadata alone also skips all of them, which is the gap to close before adoption.

Hydra stores the data, Terminal trades the chart, Designer builds the logic

The product split is wide and each part has a distinct job. Hydra loads and stores market data: many sources, high compression ratio, any data type, program access to stored data through an API, export to csv, excel, xml or database, import from csv, scheduled tasks, and auto-sync over the Internet between several Hydra instances. Terminal is the charting and trading front end, with trading from charts by clicking, arbitrary timeframes, and volume, tick, range, P&F and Renko candles plus cluster charts, box charts and Volume Profile.

Designer is the visual strategy builder, with mouse-driven strategy creation, an embedded C# editor, custom indicators, a built-in debugger and schema sharing with your own team. Shell is the ready-made graphical framework with complete C# source, strategy testing with statistics, equity and reports, saving and loading strategy settings, launching strategies in parallel, and launching them on schedule. Designer, Hydra, Terminal and the API are each labelled free.

Three READMEs and a localization generator in the build tree

Documentation is maintained in four languages at the root: README.md, README.ru.md, README.zh.md and README.ja.md, and a language switch sits at the top of the English file. The tree also carries Localization, Localization.Langs and a Localization.Generator project, plus Media, Media.Names and Media.Generator, which together suggest that user-facing strings and media names are produced by code in the build rather than edited by hand.

For a reader this cuts two ways. Translated entry points mean you are not forced to work in English.. The generator projects suggest the build may involve generated sources, so check how they run before relying on a from-scratch build.

Editorial conclusion

StockSharp offers C# trading APIs and related applications, with connector sources maintained separately. GitHub release tags are not the only distribution channel: the publisher also provides versioned NuGet packages. Verify the specific package, connector, license and venue capabilities, and test any strategy in a safe environment before using real funds.

Frequently asked questions

Official sources

1. Issues
2. Project website
3. README
4. Releases
5. StockSharp/StockSharp on GitHub