Developer-facing blueprint

Super Cut Club is a matching engine, not a job board.

In film, word of mouth is a wall that new talent cannot easily climb. This product turns professional intuition into a system: taste, resources, and availability get queried together so a director can find the right editor, not just the most connected one.

3 Data layers queried together
1 Unified match score for taste, resources, and timing

1. Core logic

The triple mapping system

The search layer needs to map three data types at once. If any one of them is missing, the shortlist feels incomplete or misleading.

Taste

Subjective creative fit

Aesthetic, mood, tempo, and visual rhythm extracted from reels instead of manually typed bios.

Tech layer CLIP embeddings + audio analysis

Resources

Objective production reality

Location, storage, workstation power, turnaround constraints, and pricing bands stored as structured filters.

Tech layer Relational data model

Availability

Real-time bookability

A strong match is only useful if the editor is actually free when the project lands.

Tech layer Availability toggles or calendar sync

Search composition

How the engine assembles a match score

Taste should lead the shortlist, but it must stay grounded by production reality and actual availability.

2. Extraction logic

Let the system read the work instead of asking for essays.

Editors should not have to over-explain themselves in a text bio if the platform can analyze their reels directly and recover visual and rhythmic signatures from the actual work.

Visual extraction

The mood sensor

Frame snapshots from a reel become vectors. That lets the system compare a director’s visual reference against what an editor actually cuts, not just what they call themselves.

OpenAI CLIP + pgvector

Rhythmic extraction

The pulse sensor

Music BPM, energy level, and pacing become machine-readable signals. High-BPM commercial editors separate naturally from slower documentary and narrative editors.

Cyanite or Sonoteller

Visual matching

CLIP turns mood into vectors

Similar-looking work begins to cluster, which makes “moody neon” or “clean documentary realism” searchable without keyword stuffing.

Rhythmic matching

Audio analysis becomes pace metadata

BPM and energy help the system separate slow-burn editors from high-tempo commercial and music-video specialists.

3. Why this exists

Discovery platform beats word of mouth.

Word of mouth is slow, biased, and geographically narrow. A matching engine gives clients a better fit while giving overlooked talent a chance to surface through the quality of the work itself.

The platform becomes useful to both sides at once: clients get a more precise shortlist, and editors outside the usual production circles get a fairer route into the room.

4. Functional blueprint

Build a capabilities engine, then wrap it in a visual-first interface.

The database should not stop at “users.” It should model practical constraints, while the front end should let reference media reshape the shortlist in real time.

Database strategy

Capabilities, not just profiles

Location and hardware

Store city, travel radius, machine strength, storage, and software stack as structured filters.

Price brackets

Use pricing tiers so basic reels and high-end commercial edits do not collapse into one vague rate field.

Availability state

Offer simple toggles first, then calendar sync later if the product proves the need.

Speed-dating interface

Reference media should rearrange the grid instantly

The UI should feel visual-first: upload a reference track or mood board, then let the ranking shift around taste, not just text filters.

Capabilities engine

Treat hardware, city, price tier, and turnaround strength as a search substrate, not profile decoration.

Reference-first interface

Let the director upload a mood board or reference track so the grid reorders around the brief’s taste profile.

Proof over proximity

The system should help a strong editor in a smaller city surface beside big-network names if the work and setup match.

Next design layer

The next useful page is the onboarding flow.

We can turn this PRD into a concrete editor onboarding sequence next: what to ask, in what order, and how to capture signal without exhausting the user.

Back to Supercut Club