Toolbar

Send To →

Opens a fresh tab for ChatGPT, Claude, Gemini, or Grok. Nothing auto-sends.

Open target after copy

When on, any “Copy …” pops the chosen model so you can paste instantly.

ADHD view

Compact maps for faster scanning in Value Miner prompts. Toggle anytime.

Transcript

Paste or Drop .txt

Raw text goes here. Timecodes like [00:12:31] are detected.

Auto Plan

We size the work into 1–6 chunks so it’s easy to mine and compile.

Break It Up

What this is

Chunk maker only. It splits the transcript into clean JSON-friendly slices. No analysis here—just tidy chunks for mining/merge.

Chunk size

Soft target for each slice. We respect timecodes and natural pauses.

Max chunks

Hard cap (1–6). Quick mode uses ≤3 to speed runs.

Manual cuts

Force cuts at exact HH:MM:SS marks when you need precision.

Buttons

🔄 Refresh chunks • 📋 Copy All Mining Prompts (parallel SM-22 mining).

Per-Chunk

📋 Copy Mining Prompt

Asks for up to 22 tight, verbatim quotes (≤280 chars) with novelty tags. Paste into your model, run, save JSONL.

📄 Copy Chunk Text

Raw text only. Handy for hand-edits or debugging.

Mine the Good Stuff

What this is

Deeper pass. It reads quotes, patterns, and tone to map people’s beliefs, trade-offs, and intent. Output is richer: maps, spotlights, and Actionable Value cards.

🎼 Copy Value Miner INIT (S1)

First map + Quote Spotlights + QuickJump + Actionable Value triplets (Big Idea → Use this → Try this).

🎼 Copy Value Miner LAYER (S2+)

Adds only new signals + fresh spotlights. Keeps the map clean.

🔥 Mind Blower (🤯/😲/🧐)

Run this after INIT/LAYER. Turns the best quotes into ADHD-friendly Actionable Value cards. Emojis = value tier: 🤯 breakthrough • 😲 surprising • 🧐 solid/practical.

📋 Miner Prompt Aggregate

One prompt to combine all segment outputs into a single worldview + final quote playlist.

🧩 Miner JSON Compiler

Converts Value Miner text into GOLD_DATASET_JSONL + MANIFEST_JSON (human-readable fields, no ranks).

Compiler

📋 Copy Merge Prompt

Paste all SM-22 JSONL outputs → returns GOLD_DATASET_JSONL, MANIFEST_JSON, and OMITTED_JSONL.

🧼 Fix JSON

Repairs messy arrays/markdown into clean JSONL (drops only the truly broken lines).

🧪 Audit Data

Checks ids, timecodes, quote length, novelty set, and basic caps/floors. Returns a crisp JSON report.

Transformers

Universal

Exec Summary, References, Applications, Dev (API/RAG). Paste GOLD_DATASET_JSONL where asked.

Synsian Set

QuickJump (10-min), Primer (20-min + blog), Decision Brief, Sales Playbook, UX Insight, PR Quote Bank, Analyst Brief, RAG Pack.

Citations

Outputs always cite [nugget_id @ tc] with verbatim quotes.

Copy & Toast

Every “Copy …” shows a green “Copied ✓”. If “Open target after copy” is on, we open your model in a new tab for immediate paste.

Timecodes

Detection

[HH:MM:SS] lines help set safe boundaries and quote timestamps.

Gaps

We prefer splits at big pauses (~20s+) or explicit starts.

Manual

Add exact cuts for surgical control.

Tokens

Estimator

Balanced chars+words model. Target is soft; 2.8k is the firm guardrail per chunk.

Cap of 6

Keeps copy-paste sane and stable across providers.

Quick mode

≤3 chunks is fast and usually enough for a solid first pass.

Workflow

1) Break It Up

Paste transcript → set Chunk size/Max/Manual cuts → chunks appear.

2) Mine

Per chunk: run 📋 Copy Mining Prompt (SM-22) or 🎼 Copy Value Miner INIT/LAYER.

3) Mind Blower

After INIT/LAYER, run Mind Blower to produce Actionable Value cards (🤯/😲/🧐).

4) Compile once

Use the right compiler (SM-22 Merge or Miner JSON) to produce GOLD_DATASET_JSONL (+ MANIFEST, + OMITTED).

5) Repair/Audit

Fix JSON if needed, then audit for confidence.

6) Transform

Turn GOLD into QuickJump, briefs, and RAG-ready packs.

Fixes

Oversized chunk

Lower target or add a manual cut, then refresh.

Messy miner JSON

Run Fix JSON → retry compile.

Duplicates

Compiler de-dupes. Reasons live in OMITTED_JSONL.

Thin tails

Planner auto-merges small ends; or nudge the target and refresh.