For production technology & pipeline

Your production's truth,
in one queryable spine.

The script, ShotGrid, FileMaker and Airtable, pulled into one governed graph your people and your agents can actually query. So "which shots use this asset and who owns the latest version" is one question, not an afternoon in Slack.

For your team:Marketing & BrandArchive & LibraryPost & VFXProduction TechnologySecurity & ITThe studio tour
The daily friction
✗The truth about this show lives in five systems and none of them agree.
✗Every 'which version is final?' turns into a Slack thread and a lost afternoon.
✗I hand-wire ShotGrid to FileMaker to Airtable, and it breaks the moment a schema moves.
What changes for you
01

The script is data before principal photography

Codex parses the screenplay into a 12-category breakdown, 9 categories auto-tagged, with cast rollups. No LLM in the parse, so the output is deterministic and diffable, not a guess you have to re-check.

02

One graph over the systems you already run

Fabric joins ShotGrid, FileMaker and Airtable into a single governed knowledge graph, mapped to MovieLabs OMC across 340 fields and 14 entity types. Permissions and time-awareness are enforced in the query itself, not bolted on after.

03

Agents query the whole stack, safely

Fabric exposes 23 entity MCP tools, so an agent platform like Claude, over MCP, the open agent-connector standard, can answer which shots use an asset and who owns the latest version, inside your permissions with a full audit trail.

04

Orchestrate ingest without re-plumbing

Nexus routes deliveries across the stack so files land classified, checksum-verified and filed in the right place once, instead of you rebuilding the hops between systems every time a delivery arrives.

Your part of the stack

The pieces that earn their keep for production technology & pipeline.

Book the ones that fit your team, or the whole platform. Each shares one index.

Codex

Turns the shooting script into a 12-category breakdown, 9 categories auto-tagged, with cast rollups, so the production document becomes structured data in week one. No LLM in the parse means the breakdown is deterministic and you can diff every revision.

Explore Codex →

Fabric

Pulls ShotGrid, FileMaker and Airtable into one governed knowledge graph, native plus 300+ more sources via the connector catalog, mapped to MovieLabs OMC across 340 fields and 14 entity types. Permissions and time-awareness live in the query, and 23 entity MCP tools open it to agents.

Explore Fabric →

Nexus

Routes and orchestrates ingest across the whole stack, so deliveries land classified, checksum-verified and filed once, instead of you maintaining brittle hand-wired hops between every system.

Explore Nexus →

Gateway

The one console over your whole media estate, cloud, on-prem and tape, with production-scoped catalog Q&A: streamed answers, permissions resolved per question, an audit row on every one, and read-only by design so a question can never change your data.

Explore Gateway →
A day in the life

One show, from script to a question anyone can ask

STEP 01

Turn the script into data

Hand Codex the screenplay. It returns a 12-category breakdown, 9 categories auto-tagged, with cast rollups, deterministically and with no LLM in the parse, so scenes, shots and cast are structured from day one.

STEP 02

Fuse the systems into one graph

Fabric joins that breakdown with ShotGrid, FileMaker and Airtable into a single OMC graph across 340 fields and 14 entity types, with permissions and version time-awareness enforced in every query.

STEP 03

Ask the whole production in plain language

Over MCP and Fabric's 23 entity tools, an agent answers 'which shots use this asset and who owns the latest version' in one pass, joining systems that never agreed before, fully audited.

Questions
How do I connect ShotGrid, FileMaker and Airtable without building a custom integration?

Fabric treats ShotGrid, FileMaker and Airtable as native sources and reaches 300+ more through its connector catalog, then maps them into one governed knowledge graph on the MovieLabs OMC model, 340 fields across 14 entity types. You query the joined graph instead of hand-wiring and maintaining point-to-point syncs that break when a schema moves.

Can an AI agent answer 'which shots use this asset and who owns the latest version'?

Yes. Fabric exposes 23 entity MCP tools over MCP, the open agent-connector standard, so an agent platform like Claude can traverse the graph that joins ShotGrid, FileMaker, Airtable and the script breakdown. Permissions and version time-awareness are enforced in the query, and every answer is audited.

Does turning the script into a breakdown rely on an LLM I cannot verify?

No. Codex parses the screenplay into a 12-category breakdown, 9 categories auto-tagged, with cast rollups, and there is no LLM in the parse. The result is deterministic, so you can diff two script revisions and see exactly what changed, rather than re-checking a model's guess each time.

How fast can I prove this on one of our real productions?

The standard pilot is four weeks on a single production. Week one Codex turns the script into breakdown data, week two indexes the archive read-only, week three runs a test ingestion pipeline through Nexus into a sandbox you name, and week four puts agents on top over MCP. You keep the context graph and a findings report in your numbers.

Prove it on your own work

Four weeks. Your team's numbers.

See the four-week pilot → Back to the studio tour