AI builds the rundown. You decide what airs.
Drop your docs. Showrunner's AI builds the rundown, wires the macros, and flags anything ambiguous before it ships. You review a draft — you don't build one.
What’s on your show today?
Paste an agenda, drop a run-of-show doc or script, or pick a starter — AI structures your show.
Files parsed in your browser. Only what you paste here leaves your device to generate the rundown.
Friday night, you're still
typing the show into a form.
The doc is already structured. You're just translating it into the timer software — line by line, cue by cue, an hour you don't get back. Showrunner's AI removes the translation step.
See the build flowStop being the data-entry layer.
The structure already lives in your run-of-show doc. Showrunner reads it the way a producer would and writes it into Tevyr surfaces directly. You review a draft. You don't build one.
See how sessions workThe script already exists.
Your producer wrote it. Your speakers signed off. The structure is right there in the doc — and it's about to get typed in by hand.
Last-minute changes break everything.
Sponsor swap at 4pm? Speaker dropped at 5? Re-typing every dependency. Linked sessions, attached scripts, scheduled polls — all rebuilt by hand.
The "new event" form is the bottleneck.
An hour per show, multiplied across every rehearsal and revision. Tevyr saves 10–30 minutes per event at the build step alone.
Every show starts as a document,
not a form nor a database.
Showrunner's AI segregates a single PDF, DOCX, or Markdown doc into the right places — session metadata into the schedule, scripts into the teleprompter, poll questions into live polling, ad-hoc cues into floating timers, sponsor mentions resolved to your sponsor library, warnings surfaced when something is ambiguous.
Get startedResolve. Don't invent.
Sponsors get matched against your library — Showrunner never invents a logo or a URL. Sessions get titled from the doc, not paraphrased. Scripts land verbatim with pacing inferred. When something is ambiguous, you get a warning, not phantom data.
See sponsor matchingScripts land in the teleprompter.
Multi-page scripts get split per session, formatting preserved, pacing inferred (WPM-aware), ready to read.
Questions become live polls.
"Show of hands" gets a single-choice poll. "Rate the keynote" gets a rating poll. Quiz blocks get a quiz, with the right answer pre-marked.
Sponsors resolved, not invented.
Showrunner matches against your sponsor library — no hallucinated logos, no fake URLs. Unknown sponsors surface as warnings, not phantom data.
Talk to it.
It edits the right rows.
"Move the keynote to 10am." "Add a 5-minute break after session 3." "Make the closing 15 minutes instead of 10." Only the rows you mention change. The rest stays untouched.
How refine worksEdits feel like edits.
Showrunner sends the prior rundown + your manual edits back to the model — so the AI knows exactly what to keep. Each save carries a unique ID, so a double-click or a flaky network never duplicates rows. Two AI models — Haiku for speed, Sonnet for accuracy — share the same usage pool.
How AI usage worksReschedule any row by talking.
"Move the keynote to 10am." Showrunner finds the row, updates the time, and re-runs the linked-to-previous chain so dependent sessions slide with it.
Rewrite a row in place.
"Replace Sponsor reel with Coffee break." The title swaps, the schedule holds, attached scripts and polls travel with the row.
Drop in what you forgot.
"Add a 5-min break after the keynote." A new row inserts where you asked, the timeline reflows, the rest of the run-of-show is untouched.
Rundown shipped.
Workflows shipped with it.
Save the rundown and Showrunner builds the macros that run it — Slack pings on session start, sponsor wall flips on session change, Notion logs on session end, lights cue on cue. One click and your event production runs itself.
See AI Macros
Build the show. Ship the show.
Showrunner builds the rundown. Macros run the rundown. Same AI, same context, same usage pool. One click jumps you from saved-rundown to a pre-loaded macro builder — no copy-paste, no re-auth, no context lost. The whole event production lives in one tool with one mental model.
Read macros docsTell your crew where to look.
Showrunner generates a Slack macro for every on_session_start. The right channel gets pinged before anyone has to ask.
The runbook writes itself.
Every on_session_end becomes a Notion row — speaker, duration, sponsor, ran/skipped status. Your post-event report exists before the lights come up.
event_logHardware cues without the cue sheet.
Showrunner reads scripted lighting calls in your doc and writes Philips Hue / WLED macros to fire them. Stage manager keeps the doc; the lights still go.
Stop typing the show into a form.
Drop the docs. Get a rundown. Run the show. Showrunner ships on every plan; quota scales with what you produce.