Writing & MediaFree
Script Coverage and Development Notes
River collects the cited moments first and rates from them, then reports the ratings its own evidence does not support.
Saltwater Cadets is an invented 112-page spec, 74 scenes, read for an invented production company. The coverage runs the usual eight categories and carries 41 cited moments, each one a scene number, a page and whether it reads for or against the rating. Five of the eight ratings survive that. Dialogue was written Good on one positive moment out of seven, and Poor is the only rating its own evidence permits. Nothing in a ticked box would have shown that.
Marketability rests on two cited moments and Craft on one, both under the three-moment floor, so they ship unrated instead of as a ticked box. Then the distribution: act three is 25% of the pages and 9.8% of the evidence, 1.43 cited moments per ten pages against act one's 5.0. Nothing there says the third act is fine. It says the coverage thinned out, which is a different problem and the reader's, not the writer's. That belongs in the coverage, stated as a limit on it.
Every template on page one is the same skeleton: header, logline, synopsis, comments, a grid, a verdict. Coverage Ink's own samples describe the standard version, a grid across 25 categories rated Excellent down to Not So Good, then Pass, Consider or Recommend for the script and for the writer. Not one of them requires a rating to name where it came from. That single requirement is what makes the grid countable, and everything else on this page follows from it.
A ticked box is not a reason, and writers can tell
The same script had been read twice before: quarter-final notes from a competition and a paid consultancy read. Between them they made 54 assertions about the script and located 2 of them, 3.7%. Both reads may have been right. Neither was usable, because a writer cannot act on "the second act sags" without knowing which pages sagged, and cannot argue with it either. The writer kept neither document. Coverage is only worth the reading behind it when the reading can be found again.
Silence gets counted too. The 41 moments touch 38 of the 74 scenes, so 36 scenes drew no comment in either direction, and those get listed. The longest uncited stretch runs from page 84 to page 90, straddling the act two break, and the last citation lands on page 104. On a script of average density, about one scene per page, a seven-page silence is five or six scenes nobody said anything about. Two runs of five pages or more went uncited, both of them inside act two.
The verdict has to reconcile with the grid it sits under. This coverage reads Pass on the script and Consider on the writer, with four of eight categories at Good. Move Dialogue to the only rating its evidence permits and it is three of eight, which is a Pass that now has a reason attached. Producer notes on a script already in development are the other direction of the same join.
How it works
Send the script
PDF or Final Draft file. Scene numbers help, and River derives them from the slug lines otherwise.
Say who reads it
A production company, a competition, a manager, or the writer. It changes what the verdict is for.
Read and cite
Every rating collects the scenes and pages that produced it before the rating is written down.
Check the grid
Ratings are tested against their own evidence, and the thin ones are reported rather than filled.
What you get
- Every rating carries the scene numbers and pages behind it, the ones for and the ones against
- Ratings contradicted by their own cited evidence, named, with the share that broke them
- Any category resting on fewer than three cited moments ships unrated rather than guessed
- Where the evidence sits by act, so a thin third act shows up as a number
- The scenes that drew no comment either way, listed, because silence is not approval
- The longest stretch of pages with nothing cited, and which act break it falls across
- Logline and synopsis written from the draft, ending included, not from the query letter
- A verdict on the script and on the writer, each naming the categories that drove it
Common questions
Is this just a template I fill in myself?
No. You get finished coverage on the script you send, with the evidence table under each rating a blank template cannot give you. What none of them supply is the rule that no rating ships without its pages. Once the verdict says go, delivering the file correctly is separate work.
What happens when the evidence disagrees with the rating?
It gets reported rather than reconciled quietly. Dialogue came out Good on one positive moment in seven, which no reading of that evidence supports, so the coverage says so and states the range the moments actually permit. You are free to overrule it, but not to leave the contradiction sitting there unremarked.
Why leave a category unrated?
Because a rating built on one cited moment is a guess with a box ticked, and it costs the credibility of the seven ratings that are real. Craft had one moment behind it and Marketability two. Both ship as unrated with the evidence shown, which is an honest position and takes nothing away from the rest of the grid.
I want notes for the writer, not a verdict for an executive.
Say so and the same evidence gets pointed at revision instead of acquisition. The scenes carrying negative moments become the work list, and the uncited scenes become the question of whether they are pulling weight. For a novel, the equivalent pass is the developmental edit brief.
Does it summarise the ending, or protect it?
It summarises it. Coverage is not marketing copy, and a synopsis that withholds the resolution is useless to the person deciding whether to option the thing. If you need a version that holds the ending back for a different reader, ask for it separately and it comes with the omission marked.
Can it read a script that is not feature length?
Yes, and the act arithmetic adjusts to the structure you tell it about. The one thing worth knowing is that scores fall off sharply outside roughly 90 to 130 pages in the largest published analysis of reader scores, so a page count outside that band gets flagged in the header rather than buried.
How is this different from feedback on a manuscript?
The mechanism is the same and the artefact is not. Coverage is written for someone deciding whether to buy, so it carries a logline, a synopsis and a verdict. Prose gets the manuscript read instead, with chapters in place of scenes and no acquisition verdict on the end. If notes already came back from a producer, production notes response prices and resolves those instead.
Script Coverage and Development Notes
Fill in the form and your workspace opens with the work already underway.