A plain-language walkthrough of what happens when you use Mucho Gusto — not the legal version. For the formal policy, see Privacy Policy and Terms.
What happens when you record a video
When you record an intro, the video itself is uploaded straight to Mux, the video hosting service that stores and streams it — Mucho Gusto's own database never holds the actual video file.
From there, a background job runs three steps on the video:
Transcription — Mux generates a text transcript of what you said. We save that transcript once; if you edit it by hand afterward, your edit always wins and is never overwritten.
Embedding — the transcript is sent to Voyage AI, which turns it into a list of numbers (a "vector embedding") that captures what the video is about. This is the fingerprint the recommendation system actually uses — more on that below.
Summary — the transcript is sent to Claude (Anthropic's AI model), which writes a short summary and a handful of topic tags for the video, shown on your profile.
Your written bio goes through the same embedding step, so teammates can be matched on what you've written even before you've recorded anything.
How your feed and recommendations work
Mucho Gusto doesn't use manual categories or tags to decide what to show you. Instead, it compares embeddings: your own "taste" is an average of the embeddings of videos you've actually watched all the way through, and the feed ranks other videos by how mathematically close their embedding is to yours. The more you watch, the more it adapts. If you're new and haven't watched anything yet, you'll see a mix of recent videos and a bit of randomness until it has something to learn from.
Your organization's Belonging Score
Your admin console shows a single Belonging Score, built from three real measurements: what share of active members have three or more real connections, what share have posted a real video, and how evenly attention (likes, comments, views) is spread across the team. No external benchmark is baked into the number — see the full methodology page for the exact weights, data sources, and what the score is not.
Gusto points, booster packs, and rarity
These are gamified nudges to get you exploring your teammates — there's no real-money mechanic anywhere in them, and an admin can dial them down (they fade automatically after your first 60 days by default, or can be turned off entirely — points keep accruing quietly either way, they just stop being shown):
You earn Gusto points for things like meeting a new teammate, liking a video, or finishing onboarding.
Booster packs unlock on a schedule (a new one from your department roughly once a day, a wider "discover" pack every few days) and open to reveal a few teammates as collectible-style cards.
Each card has a rarity tier — it's cosmetic, meant to make discovering a teammate you haven't met yet feel like a small reward, nothing more.
Your data, plainly
Your organization's data is walled off from every other organization. This is enforced at the database level, not just in application code — every table that holds your data has a security policy that Postgres itself checks on every single query, forcing your organization's ID to match before any row is even considered. There's no application code path that can accidentally see across organizations, because the database refuses the query before it gets that far.
Where things live: video files stay in Mux; your transcript, AI summary, topic tags, and embeddings live in our Postgres database, tied to that video.
If you delete your account, it disappears from the product immediately — no one can see your profile or videos anymore, effective right away. The underlying data is kept for 30 days in case you change your mind, then permanently and irreversibly erased.
You can export or delete your data yourself, right now — no request form, no waiting on an admin. Both live in Account settings: "Export My Data" downloads everything tied to your account (profile, every video you've recorded, your watch history), and the delete option starts the 30-day process above.
This page describes how the product actually works today. For the full legal policy — subprocessors, regional rights (GDPR/CCPA), and formal retention terms — see the Privacy Policy, or the security & procurement pack for a tenancy/encryption/incident-response rundown.