ScriptReader.ai is an AI script analysis tool whose core positioning is to conduct multi-reader, multi-level structured readings of complete screenplays and provide actionable directions for improvement.
[Purpose] Aimed at screenwriters and script creators, it provides feedback close to an industry coverage report before submission or revision. It emphasizes using repeated signals from multiple independent readers to increase the credibility of judgments: one reader may reflect personal taste, while the strengths and problems that repeatedly appear in the evaluations of the same complete draft by five independent readers are closer to trustworthy evidence.
[Main capabilities]
- Five reads and one synthesis: five independent readers each evaluate the complete screenplay, then a synthesis is generated, pointing out which strengths are consistently recognized, which problems repeatedly appear, and where readers disagree, and explaining why the draft cannot score higher or lower, and where rewriting should begin.
- Layered Read (patent pending): you can first look at the entire screenplay, then drill down to the act, sequence, and scene levels to observe results at different scales. The product demonstrates the analysis using known films such as Green Book, Room, and The Matrix.
- Four reading frameworks (in order: Triage, Pressure, Consequence, Clarity, but the materials explicitly state that "the other readings in the menu tell you why"):
- Triage: identify which parts need attention and which do not, corresponding to the prioritization question of "what needs to be dealt with."
- Pressure: determine "whether there is anything forcing this to happen right now."
- Consequence: determine "whether something has actually changed here, or whether it is repeating the same thing again."
- Clarity: determine "whether it is possible to understand what is happening."
- Each part is read in three ways: Design (whether it is constructed to achieve a certain purpose), Execution (whether it is well written), and Read (whether it grabs the reader). Each returns solid, soft, or broken; based on this, grade bands are divided: Keep (all three are solid), Polish (no broken, at least one soft), Rework (one broken), Two lenses low (two or more broken). The grade band reflects "how many items read weak," not how weak a single item is. Not fully read means that the part has too little evaluated material to judgeSegment, not part of the scoring range.
- Craft profile: In addition to the five-reader conclusions, each scene is scored across 14 dimensions, and the averages are compared against a ranking of more than 250 produced scripts in the library, presented as an overall shape to determine which axes are supporting the script and whether the shape reads as intuitive, conceptual, or integrated. The material emphasizes that "higher is not better, but different in shape."
- Scene-level commentary: The comprehensive report names specific scenes (for example, The Green Book scene 10, Room scene 4, The Matrix scene 31) and explains their strengths and weaknesses.
[How to use] Submit a complete script for analysis; a complete analysis is usually delivered within 1 to 2 hours. You can first try a free sample (the first three scenes, no credit card required). Each analysis report can be opened item by item to view the finished format. There is no subscription; analyses are pay-per-use or you can buy discounted credit packs.
[Applicable scenarios] Suitable for screenwriters who need structured, repeatably comparable script feedback: to see whether strengths and problems recur across multiple readers, to locate specific positions at the act, sequence, and scene level, and to compare dimensions against the library of produced scripts.
[Limitations and notes] It does not replace the real reactions of a trusted human reader to voice, subtext, and tone; it provides an AI coverage report and repeatable comparisons. Privacy is on by default, it is not used for training, and the first analysis supports a 30-day money-back guarantee. Prices start at $9.99, excluding VAT/sales tax. The material also mentions planned maintenance: temporarily unavailable from midnight to 2 a.m. Pacific Time on October 15, 2026 (07:00–09:00 UTC), with the hosting provider expecting about one hour of downtime within that window.
