Product & building

Why Helvetic Research Is an MCP Server, Not Another Dashboard

From dashboard-first workflows to a research engine built for AI-native analysis. The product logic, the workflow shift, and why MCP matters.

TRADITIONAL SAAS · WEB APP MCP SERVER · FOR AI AGENTS Claude ChatGPT Cursor MCP Server </>Run backtest Fetch market data Apply logic Analyze results

The web-app way

  • × Build the UI & dashboards
  • × Onboard every user
  • × Teach the workflow
  • × High friction

The MCP way

  • ✓ Provide tools
  • ✓ The AI handles the workflow
  • ✓ Natural-language interaction
  • ✓ Low friction
Works in Claude Web/Desktop Claude Code ChatGPT Codex Cursor OpenCode AntiGravity
The dashboard was never the product. The research engine was.
The journey

Four steps to the same realization

01

Dashboards

Built the UI, the charts, the workflows. Then spent as long teaching them as building them.

02

No-code builder

A drag-and-drop strategy builder. Powerful — but users had to learn our way of thinking in blocks.

03

AI agents grew up

By 2026, agents didn't just answer. They reason, write code, call tools, run loops, and compare outputs.

04

MCP server

Expose the engine; the agent brings the loop. Connect, prompt, and let it research.

Side by side

What actually changes

Traditional web app
Helvetic MCP
Interface
You build and maintain it
The AI you already use
Onboarding
Demos, videos, support calls
A ~5-minute connect
Workflow
Users learn "our way"
Describe ideas in plain words
You maintain
UI + workflow + education + engine
Just the engine
Friction
High
Low
For you

The AI becomes the interface. Helvetic becomes the engine.

🔌

No new platform

Stay in Claude, ChatGPT, Codex, or Cursor — the tool you already trust.

⚡

5-minute setup

Connect once with secure OAuth, then start prompting. No manual to read.

💬

Plain language

Describe the idea; the agent turns it into structured research.

🧪

A real engine

Actual backtests, robustness, significance, and reproducibility underneath.

Just ask

You describe the idea. The agent does the research.

"Backtest a 200-day moving average on bitcoin since 2020."
"Test this only when VIX is above 14."
"Compare this strategy across different stop losses."
"Find where this strategy breaks."

A research engine for AI agents. Not another dashboard.

Connect it to your AI tool, describe your idea, and let the agent research with real backtesting tools.

Start free →