// case study
StreamHalla
By the Bunshin team · June 2026
The product
StreamHalla is AI-powered VOD analytics for Twitch, YouTube, and Kick streamers. Paste a VOD link and you get back a sentiment timeline, a six-axis coaching scorecard, the exact moments viewers leaned in or dropped off, and a five-point next-stream action plan. No video is downloaded, and the first analysis is free — no account needed to start.
It exists because the native dashboards don't answer the question that actually matters. Twitch Analytics and YouTube Studio show view counts; they don't tell a streamer why the room went quiet, which moment landed, or what's changing across streams. Small streamers — the ones going live three to five times a week, trying to grow — are left guessing. StreamHalla replaces the guessing with a read on what happened and what to do about it.



What it does
A coaching scorecard, not a vanity metric. Every report grades the stream on six axes — Audience Engagement, Chat Positivity, Peak Moments, Content Consistency, Viewer Retention, and Community Vibe — and benchmarks each one against streamers at the same tier. A streamer doesn't just learn that a number went up; they learn where they're strong, where the room is slipping, and which axis to work on next.
A chat sentiment timeline that reads the room. A busy chat isn't always a good sign. StreamHalla maps the emotional arc of the whole stream and names the peaks and dips — the hype moment, the lull, the point where attention started to leak — so a streamer can rewatch the moments that mattered instead of scrubbing through hours of VOD.
A five-point next-stream plan. Each report ends with specific, encouraging actions for the next stream — not a grade left hanging, but a short list of what to change and why.
Streamer Trends — the part that compounds. Any single report is a snapshot. The real insight comes from across a streamer's whole report history: Streamer Trends finds the one variable quietly moving all their other scores — the growth lever that's rarely what you'd expect. It's the kind of pattern no single VOD, and no platform dashboard, can show.
StreamHalla works the same way across Twitch, YouTube, and Kick — one read on performance, wherever a streamer goes live.
See your own report
StreamHalla is live now. Run your first VOD analysis free at streamhalla.com — no account, no card, about 90 seconds from link to report.
Stack: React · Node.js · PostgreSQL · Dockerized on Fly.io · an LLM analysis pipeline over VOD and chat signals from Twitch, YouTube, and Kick.
Want something like this built?
Tell us what you're building. We'll figure out the rest together, no sales call, no pressure.
Start a build