Equity Deep Dive

A complete equity research session: start with deterministic analysis, follow up with questions, and pull SEC filings. This shows how deterministic and question-driven modes complement each other.

1. Search and confirm the instrument

copinance market search "Apple"

Output shows AAPL on NASDAQ. Use the exact symbol in subsequent commands.

2. Get a current quote

copinance market quote AAPL

Returns price, volume, market cap, P/E, 52-week range, and analyst target if available.

3. Run deterministic equity analysis

copinance analyze equity AAPL --timeframe mid_term

This runs the full deterministic pipeline: quote → historical prices → fundamentals → key metrics → structured report. Output is saved to .copinance/results/v2/ and displayed as Rich tables in the terminal.

Use --json for machine-readable output:

copinance analyze --json equity AAPL --timeframe mid_term | jq '.report.summary'

4. Set a profile for tailored output

If you haven’t created a profile, the CLI will prompt you. Or create one explicitly:

copinance profile create --literacy intermediate --name "Research"
copinance profile list   # note the profile ID

Pass it to analysis commands:

copinance analyze equity AAPL --timeframe mid_term --profile-id <id>

The same numbers — different narrative depth and framing per literacy level.

5. Ask a follow-up question

Switch to question-driven mode with --question. The LLM uses registered tools (quote, history, fundamentals, regime, macro) to answer:

copinance analyze equity AAPL --question "What are the main financial risks right now?"
 
copinance analyze equity AAPL --question "How does the current P/E compare to its 5-year average and sector peers?"
 
# Stream tokens as they arrive
copinance analyze --stream equity AAPL --question "Summarize the revenue trend and margin trajectory."

6. Pull historical prices for context

copinance market history AAPL --start 2025-01-01 --end 2026-01-01 --interval 1d
copinance market history AAPL --start 2025-01-01 --end 2026-01-01 --interval 1wk --limit 0

Use --no-cache to force a fresh fetch.

7. Get fundamentals detail

copinance market fundamentals AAPL --periods 5
copinance market --json fundamentals AAPL --periods 5 > aapl_fundamentals.json

8. SEC filings (requires EDGAR identity)

# Discover filings
copinance analyze equity AAPL --question "What 10-K and 10-Q filings are available and what are the accession numbers?"
 
# Read a specific filing section
copinance analyze equity AAPL --question "What does Apple say about competition risk in its latest 10-K?"
 
# Compare with another company
copinance analyze equity MSFT --question "Compare Apple's and Microsoft's revenue and net income over the last 4 quarters."

9. Machine-readable output for automation

# Full RunJobResult as JSON
copinance analyze --json equity AAPL --timeframe mid_term > aapl_result.json
 
# Extract the structured report
copinance analyze --json equity AAPL --timeframe mid_term | jq '{
  summary: .report.summary,
  metrics: .report.key_metrics,
  limitations: .report.limitations
}'

Library equivalent

import asyncio
from copinance_os.ai.llm.config import LLMConfig
from copinance_os.infra.di import get_container
from copinance_os.domain.models.job import Job, JobScope, JobTimeframe
from copinance_os.domain.models.market import MarketType
 
async def equity_analysis():
    container = get_container(
        llm_config=LLMConfig(provider="gemini", api_key="your-key", model="gemini-1.5-pro"),
        fred_api_key="your-fred-key",
    )
    orchestrator = container.research_orchestrator()
 
    job = Job(
        scope=JobScope.INSTRUMENT,
        market_type=MarketType.EQUITY,
        instrument_symbol="AAPL",
        timeframe=JobTimeframe.MID_TERM,
        execution_type="deterministic_instrument_analysis",
    )
    result = await orchestrator.run_job(job, {})
 
    if result.success and result.report:
        print(result.report.summary)
        print(result.report.key_metrics)
 
asyncio.run(equity_analysis())

For question-driven analysis via library, use AnalyzeInstrumentRequest with question and optionally conversation_history for multi-turn sessions — see Library Guide.