# Speaker Care

> Everything an accepted speaker needs, collected on time, and the day run without surprises. Enter what each accepted
> speaker was promised, the last usable date for every material, the session formats and recording intent, and who is
> travelling or speaking for the first time, and get the dated collection ladder, the one named point of contact, the
> care depth and pre-event cadence, recording consent per format with its opt-out - then the speaker packet to send and
> the day-of tech check and green-room sheet. Built on the SkillSafe skill @samber/event-speaker-experience, from
> samber/dev-event-organizer-skills (MIT), by Samuel Berthe.
> It starts after acceptance: promised terms are delivered, never renegotiated or added to. It never invents a perk, a
> fee, a date, a person, a room or a piece of equipment, and never reads anyone's identity, gender or experience from a
> name - first-time status is asked.

https://speaker-care.skillsafe.ai/

## What is free
- What the sequence position (acceptance just sent, weeks away, next week) still lets you change; the one named point of contact, with a role inbox flagged.
- The collection ladder: every material pinned to its last usable date (never the event date), the ask date, the first reminder at roughly half the remaining time and the second at the deadline; the per speaker status grid with overdue flags.
- The speaker-care depth and pre-event cadence menus re-ranked for the interview, every move named and every deleted rung struck; each speaker's own dated touches.
- Recording consent per format: the posture, the opt-out (offered, or none for a block recorded as one stream - stated in advance), and every intent-versus-told mismatch.
- The first-time speakers' care, the promised-terms ledger (delivered, escalate, where, confirm), the travel items, the day-of tech check and green-room sheet, the slide deadline with its consequence, the dated follow-up, and a verdict (name the contact first, fix first, ready to send the packet, run the day).
- Exports: the ladder (.csv and .ics), the status grid (.csv), the day-of sheet (.md), a packet draft, everything (.json).

## The metered lanes
Two tasks on gpt-terra, metered per run, routed by `task`. Input string fields: `task` (plan or packet), `event`, `speakers`, `materials`, `notes` (may be empty) and `facts`.
- `plan` replies with `summary`, `verdict`, `contact`, `ladder`, `depth`, `cadence`, `recording`, `first_timers`, `promised`, `travel` and `caveats`; the page holds every date, pick, posture and term to its own.
- `packet` replies with `packet` (title, contact, what we need by when, schedule, venue and A/V, recording per format, slide deadline, travel and expenses, on the day, after the talk), `day_of`, `thank_you` and `note`.

## Links
- API tutorial: https://speaker-care.skillsafe.ai/api.html
- Source skill: https://skillsafe.ai/skill/@samber/event-speaker-experience
- Before this: https://speaker-scout.skillsafe.ai/ (sourcing) and https://speaker-ask.skillsafe.ai/ (the invitation)