Product

FiClaw ships more than an AI chat box. It runs the quant research pipeline.

FiClaw is a quant strategy research workstation. Describe a strategy idea, and it generates code, runs a real backtest, lays out the results and compares parameter settings.

Core capabilities

AI as an operator, not just an advisor

Generic AI tools can discuss strategy logic. FiClaw executes: it writes runnable code, submits real backtests, diagnoses results and re-validates changes.

Generated strategy code

Turn a plain-language strategy idea into a structured spec and runnable Python, instead of starting from an empty file.

Real historical backtesting

FiClaw submits strategies through QuantAPI to AIQuantMind for historical backtesting. A-share history can span 10+ years, subject to data availability and strategy constraints.

Diagnosis and iteration

When results fall short, FiClaw lays out return, drawdown, win-rate and factor-decay issues so the next change can be reviewed and tested.

Parameter optimization

Extract the tunable parameter space, run grid search, and apply walk-forward checks to guard against overfitting.

What has shipped

A desktop app, cloud services, a user portal and a quant API

Current modules cover install, login, model routing, credit billing and quant integration, enough to support real trials and pilots.

FiClaw Desktop

Desktop app with first-run setup, login, built-in financial agents and a membership center. Available for macOS Apple Silicon, Windows x64 and Linux x64.

FiClawCloud

Handles authentication, credit ledger, top-up orders, model routing and call logs through an OpenAI-compatible entry point.

FiClawWeb

User portal for login, top-up, credit balance and orders, kept in sync with the desktop account system.

FiClawQuantAPI

The service layer connects FiClaw workflows to data and AIQuantMind backtesting and simulation. Live trading integration requires separately confirmed account and interface authorization.

Who it fits

Teams whose bottleneck is workflow, not a single tool

FiClaw fits teams that have tried AI for individual productivity but still struggle with how results move between research, review and execution.

Investment research and research-support teams
Quant research and strategy teams
Risk analysis and review teams
Fintech teams building a more reproducible research workflow

Next step

Want to see FiClaw on your own workflow?

Tell us about your team structure, workflow and AI goals, and we can discuss the right entry point and pilot path.

SiStarAI / AIQuantMind

Explore AIQuantMind

FiClaw organizes AI research tasks; AIQuantMind provides backtesting, simulation and authorized trading operation. Explore the platform's capabilities, integration and deployment scope.

Explore the platform →