Research and test trading strategies with Cursor

TL;DR

Use Cursor's NexusTrade connection to keep a research brief and strategy definitions beside each other in your project. Begin with a public filing read, then preserve the original rules before making one focused change. Review the diff and compare candidate and control on matching dates and assumptions. Account actions use the authenticated connection; reverting a local file does not replace reviewing the proposed portfolio operation.

Connect Cursor in the intended environment

{
  "mcpServers": {
    "nexustrade": {
      "url": "https://nexustrade.io/api/mcp"
    }
  }
}

Scope choice

The project file applies here; ~/.cursor/mcp.json applies across projects.

Merge the JSON below into .cursor/mcp.json for this project, preserving other configured servers. For all projects, use ~/.cursor/mcp.json. Open Customize, find NexusTrade and complete its browser OAuth connection.

Try a public filing read in Cursor

Run the public filing request after connecting. A fresh response may return a later filing; compare its reporting period, filing date and source URL.

Recorded NexusTrade filing reference

NexusTrade MCP sec_edgar returned this listing in Codex on 2026-10-12. Use it as a reference when checking your own connection.

Choose a separate next task

These guides are independent choices after the connection check. Review any proposed paid research, historical test or portfolio change before authorizing it.

Keep candidate and control in separate files

Save the original strategy definition before asking for a change. Review the candidate diff in Cursor, then record the test assumptions and returned operation IDs beside both definitions. A change to the exit rule should not silently change the universe or starting capital.

Research beside the strategy files

Use Cursor to work on the source-controlled research brief and portfolio definition while the NexusTrade connection provides data, historical operations and account reads. Keep the candidate’s changes visible in the editor before running it.

A useful review pairs the code change with its test period and assumptions. A generated explanation alone is not a complete strategy definition.

Review a focused diff before compute

Ask Cursor to make one concrete rule change, show the exact difference and explain what historical comparison would assess it. Preserve the original strategy and its result so the candidate does not overwrite the control.

Draft one change to this paper strategy and show the rule diff. Keep the original configuration as the control. After I approve the candidate, backtest both over the same dates with the same capital and benchmark, then explain risk, trade activity and warnings. Do not activate a live portfolio.

Use the result to decide the next experiment

Compare candidate and control on the same conditions. Inspect whether a headline return came from only a few trades or concentrated exposure. Advance to an unseen period and then a separate paper deployment when the research plan calls for it.

Account actions run through the authenticated connection. Confirm the portfolio ID in each proposed change before approving it.

Check a missing tool or failed account read

If the server is missing, check that the configuration was saved in the project or user file you intended to use. If tools appear but a read fails, complete pending OAuth and inspect the returned error. Compare the returned portfolio name and ID with NexusTrade before changing it. An empty list can be valid for a new account.

Optional: check the account before owned-record changes

Use only NexusTrade read tools. Call fetch_portfolios, then get_portfolio with one returned portfolio ID. Report the ID, name, strategies and deployment type. If the authenticated list is empty, report that result. Do not create, edit, backtest, deploy or place orders.

Continue exploring