Finlab Ai
Prototype quantitative strategies with hundreds of data columns, backtests, and example templates before committing capital or a full trading stack.
Overview
io.github.koreal6803/finlab-ai is a MCP server for the Validate phase that exposes quantitative data columns, backtesting, and strategy examples to coding agents over HTTP.
What is this MCP server?
- Quant toolkit advertised with 900+ data columns for research features
- Built-in backtesting plus 60+ strategy examples to fork and stress-test
- Remote streamable-http MCP at finlab-ai.koreal6803.workers.dev/mcp
- GitHub repository koreal6803/finlab-ai for server source
- 900+ data columns (per server description)
- 60+ strategy examples (per server description)
- Remote MCP URL on Cloudflare Workers
Community signal: 375 GitHub stars.
What problem does it solve?
You cannot tell if a trading idea is worth building when assembling data feeds and a backtester alone eats your entire weekend.
Who is it for?
Quant-curious indie builders prototyping systematic strategies with agent-assisted research before paid infra or broker APIs.
Skip if: Regulated fiduciary workflows, one-click live trading without your own risk layer, or builders who need only static chart screenshots.
What do I get? / Deliverables
After you point your agent at the remote MCP endpoint, you can iterate signals and backtests from the IDE and decide whether to invest in a full execution stack.
- Backtest runs and strategy variants explored via MCP
- Evidence whether a signal deserves a full trading integration
Recommended MCP Servers
Journey fit
Strategy ideas belong in Validate so you prove edge on historical data before Build hardens execution and Ship adds safeguards. Backtesting and column-rich datasets are prototype work—fast experiments, not production order routing.
How it compares
Hosted quant research MCP, not a brokerage integration or a generic spreadsheet skill.
Common Questions / FAQ
Who is io.github.koreal6803/finlab-ai for?
Solo builders and small quant teams who want agent-driven backtests and strategy experiments without standing up a full data lab first.
When should I use io.github.koreal6803/finlab-ai?
Use it during Validate prototype work when you need to stress-test rules on historical columns before you Build execution or Ship risk controls.
How do I add io.github.koreal6803/finlab-ai to my agent?
Register the streamable-http remote URL https://finlab-ai.koreal6803.workers.dev/mcp in your MCP client, enable remote servers, and invoke quant tools from your agent session.