Research AAPL with Claude.ai

1. Connect NexusTrade in Claude

Open Customize > Connectors, choose + Add > Add custom connector, enter NexusTrade and the endpoint below, then Continue. Review detected authentication, Continue, choose Sign in now and finish with Add. Complete NexusTrade OAuth. In this conversation, open + > Connectors and enable NexusTrade.

On Team or Enterprise, the authorized owner or role adds the connector first; each member connects their own account. Use Search and tools to select the tools needed for this task. Keep account-changing approvals enabled.

NexusTrade server URL

https://nexustrade.io/api/mcp

2. Check the returned account tools

The result is a paginated object with portfolios, total and totalPages. Walk pages when necessary. A new account may return an empty list. Keep authentication errors and missing results separate. Do not infer a successful connection from a tool list alone.

Use fetch_portfolios with the JSON below. Read a portfolioId returned by that call with get_portfolio, setting its portfolio_id argument to the same value. Compare the name, type and strategies with the intended NexusTrade account. If no portfolio exists, keep the empty result and use the company data request next. Do not create anything during this check.

{
  "include_paper": true,
  "include_live": false,
  "include_chat_portfolios": true,
  "include_positions": false,
  "limit": 20,
  "page": 1
}

3. Retrieve filings before interpreting the company

Expected output is dated filing rows or annual metric records, rather than a generic stock recommendation. If the source has no matching fiscal period, leave the comparison unresolved and follow the issuer link below.

Start the current company brief with the latest filed annual and quarterly periods returned by the tool. The historical worked example below checks your reconciliation method; it does not replace that current filing review.

Call sec_edgar with each argument object below. The first requests filing rows; the second requests inline annual XBRL metrics. Keep filing dates, accession identifiers and links returned by the tool. Do not use ingest_filings for this read-only check.

{
  "ticker": "AAPL",
  "action": "list_filings",
  "forms": [
    "10-K",
    "10-Q"
  ],
  "since": "2024-01-01",
  "limit": 10
}

{
  "ticker": "AAPL",
  "action": "fundamentals",
  "periods": "annual"
}

Worked source check: Apple FY2024

The Apple release published 2024-10-31 reports the fiscal year ended 2024-09-28. The statement values below are USD millions. This is a fixed historical source example, not the newest company report.

Dividing operating cash flow by revenue gives 30.24% for this annual period. That cash conversion ratio measures operating cash flow relative to reported sales. It is not free cash flow, a stock-return forecast or a reason to buy. A later release may restate an earlier value; retain the original source and explain any mismatch.

No matching rows. Clear the filter to see all records.

Historical issuer-reported values, not trading performance.

4. Reconcile the tool values with the actual statement

Check the cash-flow statement period: a nine-month year-to-date value cannot be compared with a twelve-month revenue value. Normalize millions versus dollars before computing a ratio. Do not add a company metric to a price rule unless its historical availability is represented.

Use the returned AAPL filing rows and annual XBRL metrics. Reconcile FY2024 revenue 391,035 and operating cash flow 118,254, both USD millions, for the year ended 2024-09-28, against this source: https://www.apple.com/newsroom/pdfs/fy2024-q4/FY24_Q4_Consolidated_Financial_Statements.pdf. Build a table with metric, raw value, unit, start/end dates, fiscal period, source publication date and source URL. Separate any quarterly, year-to-date, annual and restated records. Explain mismatches rather than choosing the most favorable number. Return an AAPL review note with the filing table, unresolved differences and returned operation IDs. Do not create, backtest, deploy or trade a portfolio.

5. Resolve one remaining question without duplicate research

If the brief needs broader evidence, review a specific research_topic message before submitting it. Preserve researchJobId from the accepted response and poll get_research_status with research_job_id set to that exact returned value. PROCESSING means poll the same job; COMPLETED supplies report/citations; FAILED supplies an error. Another research_topic call starts another billed run.

From a question to evidence

research_topic is asynchronous. Poll get_research_status with the same job ID instead of launching another billed research run for the same question.

No matching rows. Clear the filter to see all records.

Continue exploring