# Hybrid Grid

> The remote audience's half of a hybrid event, promised once and delivered session by session. Enter what the remote
> audience was promised, the published grid with which rooms can carry a feed, each session's recording consent and the
> last date a session can join the stream, and get the promise in one sentence, remote eligibility marked on the grid,
> the parity depth, a named remote crew per streamed session and the two-audience gap in the disruption order closed -
> then the stream cue track for the run of show and the remote advocate brief. Built on the SkillSafe skill
> @samber/hybrid-event-design, from samber/dev-event-organizer-skills (MIT), by Samuel Berthe. This skill is
> deliberately narrow: the format decision must already be closed (event-format-selection decides it), production
> coverage tiers belong to event-production, consent records to event-speaker-experience, the run of show to
> event-run-of-show and the budget to event-budget. It never names a streaming platform as a recommendation or a price,
> never specifies equipment, and labels its own numbers self-set.

https://hybrid-grid.skillsafe.ai/

## What is free
- The format check: if hybrid is still being weighed, "close the format decision first" and a route to event-format-selection.
- The promise in one sentence, from what the remote audience was promised and to whom.
- The remote eligibility column on the published grid: rooms that cannot carry a feed, consent no or unknown, the join/leave deadline, the remote audience's night - with every reason.
- The three menus (eligibility, parity depth, crew split) re-ranked from the interview, every deleted rung struck.
- A named remote person per streamed session, with every double-booking across concurrent streamed sessions flagged.
- The stream's cue track - pre-roll, start, Q&A handover, remote question relay, holding slate, end - as offsets from each session's start, rendered to clock times (the offsets are self-set placeholders).
- The two-audience gap in the disruption order, and who decided it; the hand-offs to production, speaker experience, run of show and budget.
- A verdict: close the format decision first, fix first, ready to hand off.
- Exports: the grid with its eligibility column (.csv), the cue track (.csv and .ics), the crew sheet (.csv), the remote notice (.md), everything (.json).

## The metered lanes
Two tasks on gpt-terra, metered per run, routed by `task`. Input string fields: `task` (design or cues), `event`, `grid`, `crew`, `notes` (may be empty) and `facts`.
- `design` replies with `summary`, `verdict`, `promise`, `eligibility`, `parity`, `crew`, `gap`, `handoffs` and `caveats`; the page holds every session, status, pick and name to its own.
- `cues` replies with the `cue_track` (every cue row copied), the `advocate_brief` (who, what to watch, the relay rule, holding, escalation), the `remote_notice` and a `note`.

## Links
- API tutorial: https://hybrid-grid.skillsafe.ai/api.html
- Source skill: https://skillsafe.ai/skill/@samber/hybrid-event-design