One desktop app, the whole quant data lifecycle
Download, view, export, serve — every step happens locally inside one app. A detailed walk-through of each module is below.
All your data in a local DuckDB file you own
Downloaded data lands in a local DuckDB file — no cloud, no SaaS. Works offline, easy to back up, and moves with your project. You keep full control of your own data.
- Single-file DuckDB — columnar storage, fast queries
- Everything stays on your machine, works offline
- Back up, copy or move it with your project
- No subscription, no cloud lock-in — the data is yours
Incrementally sync from multiple providers into local storage
Pick the interfaces you need and incrementally sync market, fundamental and macroeconomic data from several providers. Yahoo / FRED / Tushare / Binance behind one unified interface — only new data is fetched, never re-downloaded.
- Pick by interface, ticker universe or date range
- Incremental sync — only fetches what's new
- Unified interface across providers, schemas normalized
- Background tasks that don't interrupt your research
Validate completeness and correctness in-app
Browse data without context-switching to Python — charts to spot gaps, tables to spot-check fields. Fix issues before they ruin a backtest.
- Price / candlestick chart view
- Paginated tabular browsing
- Jump-to-date and ticker search
- More detail as we publish per-feature docs
Export to whatever format your scripts use
Slice and dump the local store — CSV, Excel, ready for your Python/R pipeline or to hand off to a teammate.
- CSV and Excel for analyst hand-off
- Filter-driven subset exports
- Saveable export jobs that re-run on a schedule
- Or take the whole DuckDB file into your pipeline
Plug your local data into Claude, Cursor and other AI apps via MCP
A built-in MCP server lets desktop AI assistants query your interfaces, search symbols and pull adjusted bars directly — your data never leaves the machine. Off by default, gated by a local token you control.
- One-click enable, copy-paste config for Cursor / Claude Desktop / Cline
- Honest adjustment metadata — the AI is told whether bars are adjusted
- Local-only and token-gated — off by default, you control access
- No extra install — the stdio bridge reuses the app runtime
Built-in HTTP server, called directly from your scripts
The desktop client exposes a local HTTP API. Any language that can issue an HTTP request (Python, Jupyter, R, Node) can query your local data with zero network latency.
- localhost HTTP — no network round-trip
- Language-agnostic — anything that can fetch works
- JSON responses with stable schemas
- More detail as we publish per-feature docs
Supported market data providers
Unified interface — we normalize each vendor's schema so you get clean data.
Yahoo Finance
LiveUS equities · global ETFs · crypto (no API key)
FRED
LiveUS / global macro series (free key from fred.stlouisfed.org)
Tushare
LiveChina A-share · HK · US · ETF · funds · futures — full multi-market adjustment (Tushare account)
Binance
LiveCrypto spot · USDⓈ-M & COIN-M perpetuals · funding rate (optional API key)
More providers
RoadmapAdditional providers in active integration
Looking for a specific data source? Email us: support@think-rethink.com
Want to try these out?
The free build covers the basics. Buy a license to unlock all data sources and full export / API capability.