ScriptBook = Hindsight + Foresight + Insight

Three disciplines. One patented AI. A complete decision-support system for the film industry.

6000+
screenplay parameters analyzed per script
87%
accuracy predicting financial performance
100.000+
scripts in our training & reference database
10 min
to deliver a complete 360° analysis report

How it works

Three disciplines, one AI engine

ScriptBook's patented AI is built on three interlocking scientific disciplines: data mining (Hindsight), machine learning and natural language processing (Foresight), and story feature engineering (Insight). Together they form a complete screenplay intelligence system.

Where traditional script coverage relies entirely on subjective human judgment, ScriptBook removes bias from the process by grounding every assessment in data — while preserving the creative nuance that makes great storytelling identifiable.

The result is a decision-support tool that doesn't replace human creativity or judgment — it augments it, giving producers, studios, and writers the data they need to make smarter, faster, and more confident greenlight decisions.

Data mining =

Hindsight

Hindsight is learning from past mistakes in order to build expertise. It involves critical reflection to integrate knowledge gained from historical outcomes with other knowledge sources — creating a foundation of evidence that no individual human reader can match.

In ScriptBook's context, hindsight means qualitative data mining: computational methods to gather, process, and analyze large quantities of textual data (film scripts), extracting and exploiting the patterns, structures, and signals that correlate with box office performance.

Our database contains over 100.000 scripts with known theatrical outcomes — spanning films released between 1970 and 2024. This is the largest screenplay training dataset in the industry, and the foundation of everything ScriptBook predicts.

Machine learning + natural language processing =

Foresight

Machine learning is about foresight. It comprises a set of tools that allow computers to learn by examining data about past cases with known outcomes — called "training data" or, in our context, "previously released films." Using this data, the system builds a predictive model that can then be validated on held-out cases.

Natural language processing allows ScriptBook to read, parse, and understand the semantic content of a screenplay — character dialogue, scene descriptions, narrative tone, emotional arcs — at a depth and speed no human reader can achieve in the same timeframe.

The final model can predict the commercial and critical potential of a new, unseen screenplay by comparing its signals against patterns learned from thousands of films with known real-world outcomes. This is what gives ScriptBook 87% financial accuracy.

Feature engineering =

Insight

Insight is only possible when hindsight and foresight are combined to create knowledge that can be used for present decision making. Feature engineering is the bridge between raw data and predictive power — and it is where ScriptBook invests the majority of its research effort.

Feature engineering is an art form in its own right. Like writing is an art. Like acting is an art. At ScriptBook, we engineer story-based features — the 6000+ screenplay parameters our AI measures — manually, through deep immersion in actual film scripts and careful thinking about the underlying structure of storytelling itself.

Story-based parameters cannot be automatically derived; they require human expertise combined with computational power. This means understanding the problem structure, identifying the meaningful signals in narrative data, and exposing them to predictive modeling in a form the algorithm can learn from. This is ScriptBook's deepest competitive advantage — and what our patents protect.

Put the technology to work on your script

Upload any screenplay. Get hindsight, foresight, and insight in 10 minutes.

87% financial accuracy. Patented. Trusted since 2014.

Analyse your script See the proof of concept