Verified pass. Every reader voice, its deck checked by actual card-draw evidence, and whether it is wired into the reader.
transcripts/ subdir). Both are fixed below. risingsoul and lightassefa are NOT source-starved; their content does not deal Rider-Waite pairs.| Reader | Transcripts | Passages | Pairs | of 3003 | Deck check (evidence) | State | Why |
|---|---|---|---|---|---|---|---|
| Kate | 188 | 23,370 | 2,974 | 99.0% | RWS, 23k passages | READY | Wired into read_spread via big archive. No action. |
| Chiron | 32 | 3,313 | 1,491 | 49.6% | RWS, 3.3k passages | WIRED | Big archive wired in. More transcripts to climb past 50%. |
| m3ksmith | 11 | 4,186 | 1,195 | 39.8% | RWS, 5,630 minor-card slots | WIRED ยท more tx | Combo data from a consolidated source. Pull more videos to climb. |
| eyeaphim | 51 | 1,390 | 550 | 18.3% | RWS, 35/51 tx, 1,196 draws | WIRED ยท re-extract | Genuine tarot. Spread-dump format limits pairs; re-extract for more. |
| quanyin | 50 | 790 | 453 | 15.1% | RWS, 36/50 tx, 780 draws | WIRED | Genuine tarot, 453 pairs. Fine and wired. |
| lightassefa | 49 | 12 | 9 | 0.3% | NO card draws (0) | WRONG TYPE | Channeled messages, not a card reading. Not a combo source. |
| risingsoul | 100 | 1 | 1 | 0.03% | 3 of 100 tx touch ANY draw | WRONG DECK | Quantum ORACLE deck, not RWS. No RWS pairs by design. |
TAROT_VOICES="kate,chiron" (default) or e.g. TAROT_VOICES="m3ksmith".*_combo_meanings.jsonl big archive. The same clean-combo filter applies (reversed, alt-deck, and multi-card lay-down windows rejected). build_spread_reading.py defaults to the big archives for kate + chiron. Measured ingest: kate +4,469 windows, chiron +941, plus m3ksmith / eyeaphim / quanyin readable on demand.Counts: transcript dirs under sfg/MASSIVE TAROT/<reader>/transcripts/; passages/pairs from ~/Projects/_outputs/tarot-voice/*_combo_meanings_summary.json; draw evidence from grep over each transcript corpus. "Pairs" = distinct two-card windows with real text, not co-mention. Entertainment only.