SmartQuant Product
OpenQuant
A complete SmartQuant strategy IDE with an AI-assisted path from trading idea to coding-ready specification, SmartQuant C# implementation, backtest evidence, and research review.
Overview
OpenQuant is a Windows desktop IDE for developing, researching, and running algorithmic trading strategies with SmartQuant. Codex can turn a plain-English idea into a structured strategy specification, C# implementation, backtest, and evidence-based review. You remain in control throughout the process and can inspect, edit, or run the strategy directly at any stage.
OpenQuant combines the AI-guided workflow with a familiar SmartQuant-style workspace for projects, instruments, data, properties, source code, logs, trading results, and reports. Versioning lets you explore strategy changes without losing earlier ideas, compare alternative implementations and results, and return to any previous research direction.
The Codex-assisted path from idea to evidence
Specification
Codex transforms the trading idea into a structured, implementation-ready strategy specification covering data, signals, entries, exits, position sizing, risk and money management, execution rules, simulation assumptions, parameters, and review criteria. Discuss and correct the specification before any code is generated.
Implementation
Using OpenQuant's dedicated API documentation, practical guides, and strategy templates, Codex turns the approved specification into SmartQuant C# strategy code and configuration, then builds the strategy so it is ready to run and evaluate.
Review
Codex backtests the built strategy, analyzes the structured performance report, and compares observed behavior with the specification. When evidence is insufficient, it can add targeted diagnostics, rebuild, rerun, and reanalyze within a bounded loop before reporting its conclusions and any remaining uncertainty.
From trading idea to coding-ready specification
OpenQuant turns informal strategy descriptions into explicit behavioral contracts. Codex identifies missing decisions, separates assumptions from confirmed rules, and organizes the result so implementation and later research review use the same source of truth.
Market and data definition
Defines the objective, instrument universe, required market data, bar construction, sessions, and data availability assumptions.
Trading behavior
Makes signals, entries, exits, position management, sizing, money management, and risk controls explicit.
Execution and simulation
Documents order behavior, execution and simulation assumptions, configured costs, parameters, and unresolved feasibility questions without inventing missing platform details.
Research and acceptance criteria
Defines what the backtest must report, how behavior will be reviewed, and which outcomes determine whether the strategy meets its objective.
Assumptions and open questions
Separates confirmed trading rules from unresolved decisions so missing details are visible before implementation begins.
Shared source of truth
The approved specification becomes the contract used by Codex during implementation and the baseline for evidence-based Review.
SmartQuant knowledge built for Codex
OpenQuant gives Codex curated, local knowledge of the SmartQuant development model instead of relying on generic C# knowledge or guessed APIs. Dedicated references and practical guidance help Codex produce implementations that fit the platform and build successfully.
Dedicated API references and guides
Local SmartQuant API documentation, implementation quickstarts, and focused guides explain strategy events, indicators, positions, accounts, exchanges, configuration, and other platform behavior.
SmartQuant strategy templates
Approved strategy and workspace templates give Codex reliable starting structures and help it select the appropriate SmartQuant architecture for each strategy.
Specification-driven implementation
The approved strategy specification remains the implementation source of truth, keeping generated source and configuration aligned with reviewed trading behavior.
Controlled build feedback
Standardized build tools and compiler output give Codex a controlled feedback loop for repairing implementation errors within explicit retry limits.
Iterative AI research review
OpenQuant does more than summarize backtest statistics. Codex investigates whether the strategy behaved as specified, identifies missing evidence, and adds compact diagnostic instrumentation when the default report cannot support a defensible conclusion.
Evidence-driven diagnostic loop
When required, Codex adds targeted observability, rebuilds the strategy, repeats the backtest, and reanalyzes the results within an explicit iteration limit.
Research without silent strategy changes
The Review phase may improve diagnostics, but it never changes trading mechanics, risk rules, parameters, or simulator assumptions and does not perform autonomous performance tuning.
The result is a research report grounded in observable behavior. It states defensible conclusions, identifies unresolved uncertainty when evidence remains insufficient, and provides separate actionable feedback for the next strategy-specification or diagnostics version.
A complete strategy IDE, with or without Codex
OpenQuant combines AI-assisted development with full user control. Every specification, implementation, configuration, and result remains available for direct inspection, editing, validation, and operation without depending on Codex.
Transparent AI assistance
Codex works against visible specifications, source, configuration, results, and discussion history. Users can inspect, challenge, or revise every AI-produced decision and artifact.
Independent hands-on control
Developers can take over at any stage to edit, run, validate, and operate the strategy directly, using Codex where it adds value rather than making it a requirement.
One strategy from research to operation
Use the same SmartQuant strategy project across backtesting, paper trading, and live operation with the appropriate providers, connectivity, and runtime configuration.
Unified trading-result analysis
Understand how the strategy develops as it trades by examining market context, strategy logs, charts, orders, fills, positions, account value, PnL, and equity together.
Users can move freely between AI-assisted research and conventional hands-on development while keeping one continuous, inspectable strategy history.
Versioned research, not disposable prompts
Every OpenQuant version is a reproducible development checkpoint. It connects the current strategy specification with its generated code, configuration, build output, backtest output, and review.
Each version preserves the complete research chain: the approved trading rules, implemented strategy, backtest evidence, Codex review, and recommendations for the next iteration. This makes it possible to understand not only what changed, but why it changed and what evidence supported the decision.
- Create a new version before exploring a change or implementing a revised specification.
- Preserve discussion transcripts, generated source, configuration, logs, reports, and feedback together.
- Reuse a prior implementation as a baseline while keeping the latest specification as the behavioral target.
- Apply evidence-based Review feedback to create the next strategy-specification version.
Built on SmartQuant
OpenQuant develops event-driven strategies for the SmartQuant platform. Its curated documentation, templates, editors, and runtime tools cover SmartQuant strategy architecture, configuration-driven execution, instruments, market data, event logs, and backtest reporting.
The application can work with supporting SmartQuant infrastructure such as QuantController and QuantBase. Standardized tools keep automated builds and research backtests controlled and repeatable, while the IDE also supports direct strategy operation.
What OpenQuant delivers
- • A clear path from a plain-English trading idea to an executable SmartQuant strategy.
- • A structured collaboration loop between the user, Codex, and controlled local tools.
- • Reproducible strategy versions with complete, inspectable research artifacts.
- • Evidence-based feedback for improving the next strategy version.
- • A hands-on SmartQuant IDE for editing, running, and inspecting strategies without Codex.
OpenQuant brings specification, implementation, research review, trading-result inspection, and strategy operation together in one focused environment.