Earnings Call Processor v1.1.0

Fully automated pipeline: audio → transcript → analysis + stock data → Feishu draft.

Quick Start (One Command)

python3 scripts/process_call.py --ticker AAPL --quarter "Q1 2025" --audio call.mp3

This outputs structured JSON. The agent then automatically:

1. Fetches stock data via finance skill
2. Applies analysis framework → structured markdown
3. Creates Feishu draft via feishu-doc skill

Workflow Details

Step 1: Input

Accept one of:

• Audio file (mp3/wav/m4a) → transcribe via scripts/transcribe.py
• Transcript file (txt/md) → use directly
• Ticker + quarter → fetch from public sources

Step 2: Transcription (audio only)

python3 scripts/transcribe.py <audio_file> [--model base] [--language en] [--output out.txt]

Uses faster-whisper (CPU, int8 quantized). Model sizes: tiny/base/small/medium/large.

Step 3: Stock Data Enrichment

Call these finance skill tools:

• yfinance_get_stock_price(symbol) — current price
• yfinance_get_historical_data(symbol, period="5d") — earnings window price action
• yfinance_get_stock_info(symbol) — market cap, PE, sector, industry

Step 4: Structured Analysis

Read references/analysis_framework.md for the full template. Required sections:

1. 执行摘要 — beat/miss/meet, headline numbers, market reaction
2. 核心财务指标 — revenue, EPS, margins with YoY and vs-consensus
3. 管理层基调 — sentiment classification with quoted evidence
4. 战略亮点 — M&A, products, expansion, capital allocation
5. Q&A 重点 — top 3-5 analyst questions + management responses
6. 前瞻指引 — raised/lowered/maintained with details
7. 股价关联 — after-hours reaction, volume, sector comparison
8. 投资要点 — bull/bear/risk/watch items

Step 5: Feishu Draft

python3 .claude/skills/feishu-doc/scripts/doc_ctl.py create "财报分析: {COMPANY} {QUARTER}" --content "<structured markdown>"

Title format: 财报分析: {COMPANY_NAME} {YEAR} Q{N}

Step 6: Auto-Notify (v1.1.0 new)

After Feishu draft creation, send a brief summary message to the user channel with doc link.

Configuration

v1.1.0 Changelog

• Added scripts/process_call.py end-to-end pipeline script
• Added auto-notify after Feishu draft creation
• Added version field to frontmatter
• Improved analysis framework with investment takeaways section
• Added quick start one-command workflow