who it's for
itsgiving is for directors who make music videos and short films, and for the people who commission them. They share one problem: the work has to land with an audience the director has never met.
- Music-video directors
You're bidding for a video, often against two other directors, with a few days and no research budget. You know your own references. The artist's fans are the unknown.
A treatment built on what those fans love: the films, the clothes, the places, and the part of the city where they are.
- Short-film makers
You know the films your short is after. You need a lookbook that gets your cinematographer, production designer and costume on the same page, and convinces a festival, a grant or a crowdfunder.
A lookbook that reads the audience of those films and turns it into references, wardrobe and real places to scout.
- Labels, artist managers and producers
You brief directors and choose between their treatments. You need to know which one speaks to the artist's audience, not only to the director's taste.
Treatments where every film, face, brand and place shows Qloo's receipt: how it ranks with that audience, and what put it there.
What it is
A treatment agent. Give it an artist, or the films a short is after, and a shoot city. In about a minute it writes a treatment you can pitch, built on what that audience loves.
Add the films you love and it finds where your taste meets theirs: your voice, their taste. Every treatment prints to PDF and takes a rework in one tap: darker, more 90s, deeper cuts, cheaper locations. No account is needed to build one; sign in with Google or GitHub to save it and share it by link.
Each treatment holds
- A title, a logline and a concept
- The tone, in the audience's own words
- Reference films, and what to borrow from each
- Beats, shots, light and a palette
- Wardrobe brands the fans favour
- Real locations around a shoot base, on a map
What Qloo is
Qloo is a taste graph: a map of how people's tastes connect across culture. It calls itself the privacy-first cultural intelligence layer for AI, and gives agents something a language model can't: what real audiences love, and where they are.
- Across culture. Music artists, films and TV, books, podcasts, brands, places and neighbourhoods: over 250 million entries, by the hackathon's count.
- Across domains. Its signature is cross-domain affinity. Start from the fans of one artist and it ranks the films, clothes and places they over-index on.
- Down to the street. It maps where an audience's taste is strongest in a city, tile by tile, so a shoot can be based where the fans are.
- Without identity. It learns from anonymized signals and needs no names, emails or device IDs. What it says describes groups, never a person.
More at qloo.com.
How itsgiving uses Qloo
Code makes every Qloo call. A model plans and writes, but only from what Qloo returned, so nothing in a treatment is invented. A treatment takes about 20 Qloo calls and two model calls.
- Step 1
Find the names in Qloo
The artist, the films and the shoot city become Qloo entries. A name that could mean two titles is asked about, never guessed. Musicians Qloo files under people are found there, and an artist it doesn't have can be read through stand-in artists with a similar audience.
Qloo calls
/search
/entities
- Step 2
Read the audience
What the fans over-index on, in Qloo's own words (moods, genres, styles), the age and gender they lean to, and the film subgenres they love, which suggest an era look for the cover. Beside it, the actors they know, as faces for the story's characters, and the artists they play, as a short film's music.
Qloo calls
/v2/insights, taste tags
/v2/insights, demographics
/v2/insights, film subgenres
/v2/insights, actors
/v2/insights, artists
- Step 3
Compare with your taste
Add the films you love and Qloo compares them with the audience, and ranks each of your films for it. What you share becomes the spine of the treatment: your voice, their taste.
Qloo calls
/v2/analysis/compare
/v2/insights, your films scored
- Step 4
Map where the fans are
Qloo's heatmap scores the shoot city tile by tile for this audience, and itsgiving names the three strongest areas.
Qloo calls
/v2/insights, heatmap
/search, areas nearby
- Step 5
Plan the creative direction
A model picks the direction: the tone from Qloo's tags, a few style words, and which area to base the shoot in. It can only choose what Qloo returned.
The model chooses, from Qloo's list.
- Step 6
Find references, wardrobe and places
Films and series the audience over-indexes on, from the canon to deep cuts; fashion brands steered by the style words, with their style and prices; and real places within 2 km of the base, inside the budget, with photos, opening hours and the venue's own contact.
Qloo calls
/v2/tags, style words
/v2/insights, films and series
/v2/insights, brands
/v2/insights, places
- Step 7
Write the treatment
The model writes the treatment and may cite only what Qloo returned. It sketches the story's characters, each drawn from a character in one of the films or series, as TMDB credits it. Every film, face, brand and place keeps its receipt: its rank for this audience, and its affinity.
The model writes, citing only Qloo's picks.
What a treatment can't tell you
- Qloo describes groups. A treatment says what an audience tends to love, never what one person will do.
- Qloo scores feature films, series and artists, not short films or songs. A short's audience comes from the feature films it's after, and a video's from the artist, not the track. A song named in a treatment comes from Qloo's list of the artist's notable songs: clear it before you use it.
- Characters are invented for the story. The characters they're drawn from, the parts an actor is known for and the photo of a face Qloo has no picture of come from TMDB, and the actors are faces this audience knows, for the presence a part needs: whether anyone is free or interested is for you to ask.
- Not every artist is in Qloo. When one is missing, stand-in artists carry the audience, and the treatment says so.
- Locations are a scout list. Availability and filming permission are yours to confirm with each venue.
Accounts and your data
Building a treatment needs no account. Signing in with Google or GitHub is only for saving treatments and sharing them. itsgiving keeps your name and email from that sign-in and the treatments you save; none of it goes to Qloo or the AI models.
A saved treatment is private until you turn on its share link. Shared pages are never indexed and don't say whose they are, and deleting a treatment takes its link down with it.
Deleting your account, at the foot of Your treatments, removes everything you saved and every link to it. Your Google or GitHub account stays as it was.
Make one for your next video or short.
An artist or a film, and a city, are enough to start.