Earnings Call LLM - bpleone / trade

📄 Transcript

Auto-pulls the latest available transcript by ticker — free, no key.

Sourced via your worker from Motley Fool. Fresh calls can lag a few days after the print; the fetched date is shown. Or paste / drop a transcript below.

Drop a .txt, .md or .pdf here, or click to choose — or just paste below

0 chars

🎯 Training Scoreboard

Calls

Labeled

—

Agree ±1

—

Outcome hit

How it learns: each read you label (your score, a note, the realized move) is folded into a calibration profile that's injected into every future prompt. It's prompt-level continual learning — not weight fine-tuning — so it converges to your read over time.

📝 Standing Rules

One rule per line. These are sent on every analysis (e.g. "MGMT sandbags Q4 guidance", "discount buyback hype", "weight datacenter capex above EPS beats").

🧠 What it has learned about you

The exact calibration text injected into every analysis prompt, built from your labels, rules, and outcomes. This is the "training."

📊

Paste a transcript and hit Analyze. The read shows up here — then label it to train the model.

Methodology & disclaimer. Each read is generated by Claude (opus-4-8) from the earnings-call transcript you supply, weighted toward guidance-vs-consensus, management tone, and demand signals; bull and bear points are quoted verbatim and figures are never invented. This is analysis and research tooling — not investment advice, not a recommendation, and not a substitute for the company’s official filings and disclosures. Transcripts, including any auto-fetched text, are third-party sourced and may lag or contain errors; verify against primary sources before acting.