README.md

Daily Trades

Daily market intelligence newsletter that surfaces structural trends — slow-moving, under-the-radar shifts — and maps them to public companies before mainstream coverage catches up.

How It Works

AI pipeline runs daily: ingests RSS headlines, filters for structural trends, scores them, and writes articles with company picks. One article per day goes out via Beehiiv.

Tech Stack

LLM: Anthropic API — Haiku 4.5 (filter), Opus 4.6 (score + write)
Database: SQLite (data/daily_trades.db)
RSS: feedparser (16 active feeds)
Runtime: Python 3.12 on homelab server
Review UI: Web dashboard on port 8080

Pipeline


python3 scripts/run.py     # Full daily pipeline:
                           #   1. Ingest RSS feeds
                           #   2. Filter + rank (Haiku)
                           #   3. Apply time decay
                           #   4. Score all filtered (Opus)
                           #   5. Write top 3 articles (Opus)

See pipeline.md for architecture details.

Commands


python3 scripts/run.py              # Full pipeline
python3 scripts/run.py --ingest     # Fetch RSS feeds
python3 scripts/run.py --filter     # Filter + rank new headlines
python3 scripts/run.py --decay      # Apply time decay
python3 scripts/run.py --score      # Score filtered headlines
python3 scripts/run.py --write      # Write top 3 articles
python3 scripts/run.py --top        # Show top scored headlines
python3 scripts/run.py --pool       # Pool health check
python3 scripts/run.py --sources    # List feeds
python3 scripts/run.py --serve      # Dashboard on :8080

Key Files

| File | What |

|---|---|

| product.md | Product definition, audience, comms strategy (source of truth) |

| pipeline.md | Pipeline architecture and daily flow |

| example-format.md | 5 example issues showing target output |

| mfm-style-guide.md | Shaan Puri voice analysis driving the writing tone |

| prompts/filter.md | Filter prompt (loaded at runtime) |

| prompts/score.md | Scoring prompt with ticker rules |

| prompts/write.md | Writing prompt with style and structure rules |

| scripts/run.py | Main pipeline |

| scripts/init_db.py | Database schema |

Status

v0.3 — Context-aware pipeline with Opus scoring/writing. Trend and ticker dedup. Issue tracking. Dashboard with manual "mark used" workflow.