Celestia Intelligence
Content keeps growing.Attention is harder to win.
Celestia measures what people say and do around an idea, then turns it into a clear read on what to make, where to launch it, and who will show up for it. It reads the same way for a studio slate as for a creator’s channel.
Domestic-power themes resonate loudly with thriller audiences; comp validity and talent momentum are the weak legs — refine packaging before greenlight math.
Signal-strength reading of external/audience-side data. The producer makes the creative Develop / Refine / Pass call.
How much to trust the read above — a gauge on the analysis, not part of it.
Executive summary
Audience chatter around domestic-worker power dynamics is loud, current, and majority-positivec1c2. The concept's genre lane — the female-led domestic thriller — is rising on both behavioral and conversational signalc5.
The main liability is the comp slate: two of the four anchors are prestige outliers whose audience profile does not transfer to this packagec6c7. Talent momentum is flat rather than falling, which reads as neutral for packagingc8.
- HClass-divide topic outperforms the genre baseline on volume and sentimentc1c3
- MFemale-led thriller momentum rising on comp-genre trending and audience engagementc5
- HComp 'Parasite' flags weak-fit: prestige outlier, not a scale modelc6
Primary risk
- HComp slate over-indexes on prestige outliers; financial translation unreliablec6c7
Primary opportunity
- MGaslighting-in-the-home vector shows unserved appetite in thriller audiencesc2
Money moves first. Proof shows up last.
A film commits its budget before there's any real evidence an audience exists. A creator commits weeks of production time and finds out from the analytics, after the video is live. Different pipelines, same blind spot. Celestia puts the evidence in the room while it can still change the decision.
The wrong things get funded
Projects move forward before clear audience and market evidence is available.
Strong markets get missed
Demand, cultural shifts and audience momentum are spotted too late to act on.
Marketing money gets wasted
Campaigns go broad without knowing who is genuinely interested or which message will land.
What people say. What people do.
Two kinds of evidence, read together. One explains the attention; the other shows where it's actually going.
Stated
Posts, comments, reviews, threads, press. Something a person chose to write. It explains the attention.
- TikTok
- YouTube
- X (Twitter)
- Bluesky
- Steam
- Twitch
- GDELT
- Hacker News
Observed
Searches, views, tickets, ratings. Nobody is asked anything; behaviour is simply counted.
- TMDB
- Wikipedia
- Box Office Mojo
- RT / Metacritic
- Google Trends
- Google Ngrams
Default weighting
Adjusted per project
30Say
Do70
Behaviour carries more weight because it sits closer to real attention. This is a starting weight, not a fixed rule. It moves with the project, the market and how much data actually came back.
Ask about an idea.Get an answer you can check.
Describe the concept, attach the script if there is one, name the comps. Back comes one connected read, and every claim in it links to the source it came from.
Concept input
Examples
- Sinners (2025)
- The Housemaid (2025)
- Anora (2024)
- The Smashing Machine (2025)
- Project Hail Mary (2026)
The idea
01Title
Logline
The script
02Drop script or browse
PDF · DOCX · Fountain · TXT · MD — max 5MB
Market context
03Genre
The idea, the script, the comps.
A logline is enough to start. Attach the screenplay and it reads the themes out of the script itself, then goes looking for the audience those themes already have.
Scripts are processed server-side, never used to train a model, and deleted after 24 hours.
Audience pulse by topic
4 topics from script briefSentiment across topics
- positive1
- mixed1
- negative1
- unclear1
Momentum across topics
- rising2
- stable1
- unknown1
Class divide and exploitation of trust in domestic workers
theme- HViewers frame the tension as 'invisible abuse — shift work as a weapon'c1
- MStrong-cast requirement is the recurring caveat in positive chatterc2
- MMomentum concentrated on YouTube and Steam community; GDELT quietc3
How we measured — 2 vectors · 3 anchor works · filter kept 35/47
Gaslighting and psychological manipulation inside the home
theme- MCommenters share first-person recognition stories; anger is aimed at abusers, not at the filmsc2
- LRecurring praise for stories that show the slow accumulation instead of a twist revealc4
Thin sample after filtering — read as directional, not conclusive.
How we measured — filter kept 23/32
Comp reality check
financial translation of the audience readThe Housemaid (2010)
ValidStructural comp: same engine (housekeeper POV + intimate threat), comparable scale.
Silent
- MSustained Wikipedia pageviews and TMDB popularity years after releasec9
Loud
- MYouTube reaction videos still actively compare new domestic thrillers against itc3
Parasite (2019)
Weak fitAspirational, not structural — usable as a thematic reference, not for greenlight math.
Silent
- HBox office and awards profile driven by a once-a-decade prestige wavec6
Loud
- MAudience reviews center the director brand more than the domestic-thriller mechanicsc7
Coverage & sources
how representative this read is — raw signal volume by platform- Topics analyzed:
- 4
- Returned signal:
- 4 / 4
- Loud-signal items:
- 958
- Silent-signal items:
- 11
- Tool calls:
- 43
Signal volume by platform
- YouTube comments472
- TikTok comments183
- Steam community100
- TikTok63
- Steam thread replies63
- TMDB reviews52
- Press (GDELT / Google News)25
- Wikipedia pageviews11
- X0
- X replies0
- Twitch0
- Google Trends0
By platform
Three decisions. One read.
What to make, where to launch it, who to reach. Each one changes the other two, which is why Celestia treats them as a single question, whether the project is a feature film or next month's flagship video.
- Greenlight01
What should we make?
Whether the audience for this idea already exists, and how warm it is.
- Distribution02
Where and when should it launch?
Which markets and moments the demand is actually concentrated in.
- Marketing03
Who will care, and how do we reach them?
Which audiences to spend against, and which angle they respond to.
The questions are the same, whoever's asking.
Celestia starts in film, where each decision carries the most money. The same engine reads for anyone whose next release is a bet on an audience.
Studios & funds
Slate and greenlight calls with the comp evidence checked instead of assumed, and a paper trail for every claim in the room.
Producers & distributors
Where the demand for a title actually sits, which markets are warm, and when a release meets the audience instead of chasing it.
Creators & creator studios
YouTubers, podcasters, writers. What to make next, when to drop it, and which angle your audience is already asking for.
See what the audience already knows.
Bring a project you’re weighing right now and we’ll run it live.