Softrun lets speakers, comedians, and writers test-run a script against a synthetic audience before they ever step on stage. Upload a keynote, TED talk, sales pitch, wedding speech, eulogy, or stand-up set, build the room you'll actually be facing, and get honest, per-person reactions in minutes. No need to gather beta listeners or wait for the live moment to find out what lands.
You start by building an audience of 1 to 100 AI personas, tuned to the room you expect: gender split, age range, countries, occupations, education, income, interests, and personality or humor style. Each persona comes with its own name, bio, and avatar. Then you drop in your script — paste it, or upload a .docx, .pdf, or .txt — and run it. Every persona reacts in character with a reaction tag (laughed, smirked, didn't land, tuned out for comedy; engaged, nodded, drifted for talks), a 0–10 score, the lines they loved, their criticisms, and concrete suggestions. Softrun rolls all of it into an aggregated report of strengths, weaknesses, recommendations, and exactly which lines landed versus bombed.
Because audiences and scripts are reusable, Softrun is built for iteration: compare runs across drafts, dial the feedback tone from friendly to mixed to critical to brutal, and reuse the same audience across multiple scripts to see how different material plays to the same room. Purpose-specific modes for stand-up, general speeches, keynotes, TED talks, pitches, wedding speeches, and eulogies each apply their own reaction taxonomy and framing, so the feedback fits the format.
Whether you're rehearsing a conference keynote, tightening a comedy set, sharpening an investor pitch, or making sure a wedding speech or eulogy hits the right note, Softrun gives you a room full of reactions before it counts. Plans range from a free tier for trying it out to team plans with larger audiences, more runs, and unlimited history.
