The short answer
Paste your first page into the checker above and you will get a score out of 100 plus a six-item checklist that marks each common rejection signal as pass or fail. The six checks cover the length of your opening line, whether voice or action appears early, how many capitalized names crowd the page, how often filtering and hedging words appear, whether backstory stalls the present scene, and whether the page opens with a dream. A page that passes five or six checks with a score of 75 or higher is ready to carry a query batch. A page that passes three or four needs one focused revision pass. A page that passes two or fewer, or scores under 40, should be rewritten before it goes to any agent.
That is the whole method. The tool does not judge your concept, your prose style, or your chances with any particular agent. It counts concrete features of the text you paste and compares them against limits drawn from what agents publicly cite as fast rejection reasons. Use the score to decide what to fix, fix the top miss first, then run the page again.
Why page one decides so much
An agent reading queries faces a stack of hundreds of submissions. The query letter earns thirty seconds. The first page earns perhaps two minutes, and most of that decision happens in the first three paragraphs. This is not because agents are careless. It is because the same small set of craft problems appears again and again in rejected openings, and experienced readers learn to spot them fast.
When an agent says a submission was rejected on the strength of page one, the cause is almost always one of a short list: the opening line is long and shapeless, nothing happens and nobody speaks, six unfamiliar names arrive before the reader has a reason to remember any of them, every perception arrives wrapped in felt and saw and suddenly, the first scene stops so the narrator can deliver family history, or the whole page turns out to be a dream. Each of these tells the agent the manuscript will need heavy editorial work, and the agent moves on.
Notice what is missing from that list. Agents rarely reject on page one because the premise is unfamiliar, because the genre is crowded, or because the voice is unusual. Unusual voices get requested. Crowded genres get requested when the opening promises a fresh version. What gets rejected fast is craft friction: anything that makes the first 300 words harder to read than they need to be. That is good news, because craft friction is fixable in an afternoon, while a weak concept takes months to rethink.
What the six checks look for
The checklist on the results screen tests six specific signals. Each one maps to a complaint agents raise in interviews, conference panels, manuscript wish lists, and rejection letters. Treat them as screening signals, not as laws of fiction. Published novels break every one of them. The point is that a debut manuscript breaking them carries a cost, and you should know the cost before you pay it.
First, the opening line must run 25 words or fewer. Long opening lines correlate with throat-clearing: the writer circling the scene instead of entering it. A short first line forces a choice about who is present and what is happening.
Second, voice or action must appear on the page. That means the page opens with an action or a line of dialogue, or quoted speech appears somewhere in the pasted text. Readers bond with a speaking, moving character faster than with a described room.
Third, capitalized names must stay under a per-genre limit. New names are memory work, and a reader on page one has no reason yet to do that work gladly. Fantasy and science fiction get a higher limit because invented places and titles are part of the genre contract.
Fourth, filtering and hedging words must stay at three or fewer. Words such as felt, saw, seemed, suddenly, just, really, and very place a pane of glass between the reader and the event. The reader wants the cold water, not a report that the narrator felt the cold water.
Fifth, backstory flags must be zero. Phrases pointing at childhood, years ago, births, and the way things were in those days signal that the present scene is about to pause for history. History matters, but page one belongs to the present.
Sixth, the page must not open with a dream. Waking-up and dream openings are among the most frequently named agent turn-offs, because they spend the reader's first investment of attention on events that did not happen.
How the scoring math works
The tool starts every page at 100 points and subtracts fixed penalties. Filtering and hedging hits cost 5 points each, up to a maximum of 30. Tell-style words such as was, were, told, and knew cost 3 points each, up to 20. Backstory flags cost 12 points each, up to 36. An opening line over 25 words costs 10. A dream opening costs 15. A page with no quoted speech that opens in description, thought, or dream mode loses 8 more. Name density over the genre limit costs 8. The result is clamped between 0 and 100.
The verdict comes from two numbers together: how many of the six checks pass, and the score. Pass five or six checks with a score of 75 or above and the verdict is a strong page. Pass two or fewer, or score below 40, and the verdict calls for a rewrite. Everything in between is a middle score, which means the page is close and usually needs exactly one targeted fix.
Here is a worked example with real arithmetic. Suppose you paste 240 words of a thriller that opens in action mode. The opening line runs 14 words, so the line check passes and no 10-point deduction applies. The text contains one line of quoted speech, so the voice check passes and no 8-point deduction applies. The page names two people, against a limit of 8 for thrillers at this length, so the name check passes. The text contains the words felt once and really once, which is 2 filtering hits, so the filter check passes and the deduction is 2 times 5, or 10 points. One was appears, which is 1 tell word for 3 points. No backstory phrases appear. The page does not open with a dream. The score is 100 minus 10 minus 3, which is 87, with all six checks passing. The verdict is a strong page, and the recommendation is to read the page aloud and polish rhythm.
Now change one fact. Suppose the same page contains suddenly twice more and just once more, bringing filtering hits to 5. The filter check fails, and the deduction becomes 5 times 5, or 25 points. The score falls to 100 minus 25 minus 3, which is 72. Five checks pass but the score sits below 75, so the verdict drops to the middle band, and the top recommendation names the exact number of filtering hits to remove. That is the whole system: counts in, deductions out, and the fix list ordered by what costs the most points.
Illustrative example one: a thriller opening that passes
The following hypothetical passage is invented to show what a passing page looks like. It is illustrative, not a model to imitate word for word.
Run. Mara took the stairs two at a time with the envelope flat against her ribs. A door slammed one floor below. She stopped breathing and listened to shoes on concrete, unhurried, coming up. The gray car would be double-parked on Mercer with Luis behind the wheel, if Luis had made the light on Canal. She pushed through the roof door into white daylight and wind. The river glittered beyond the railing. Her phone buzzed once against her palm. She did not look at it. Below, the shoes kept climbing.
Walk through the scoring. The opening line is a single word, so the line check passes easily. The page opens with action, a woman fleeing with an envelope, and quoted speech is not even needed because the opening mode already satisfies the voice check. Two capitalized names appear, Mara and Luis, plus two street names, Mercer and Canal, for a small total well under the thriller limit. Filtering words are absent: nothing is felt, seen, or suddenly done. No backstory phrases appear. Nobody wakes from a dream. The score lands in the high eighties or nineties depending on the handful of was-type words, and all six checks pass.
Notice what the passage does not do. It never explains who Mara is, who pursues her, what the envelope holds, or why Luis matters. New writers often load those answers onto page one from fear the reader will be confused. The passage trusts confusion in small doses. A reader tolerates four unknowns when a woman is running. That tolerance is the engine of commercial openings, and it costs nothing to build: remove explanation, keep motion, and let the questions pull the reader down the page.
Hypothetical example two: a literary opening that fails three checks
This second hypothetical passage is also invented. It shows how a quiet, carefully written page can still trip the most common signals.
She woke with a start from the dream, which felt so very real that she just lay there. She was born in the valley where her mother grew up, and in those days the winters were long. Her childhood home had burned years ago, and she suddenly remembered it all. The ceiling needed paint. The house kept the cold the way some people keep grudges, and she wondered whether selling it would count as mercy or theft.
The writing is grammatical and the images are decent. The score still collapses. The opening mode is dream, which fails the sixth check and deducts 15 points. Backstory flags appear four times: born, grew up, in those days, childhood, years ago is five by the tool's phrase list, which fails the fifth check and deducts the maximum against the backstory cap. Filtering hits pile up: felt, very, just, suddenly, wondered, really is absent but the count still clears the limit of 3 several times over, failing the fourth check and deducting 25 to 30 points. Tell words add was, were, and had forms. The final score lands near the rewrite band with only two or three checks passing.
The repair preserves everything good. Delete the dream sentence and start with the ceiling that needs paint. Move the valley, the mother, the burned house, and the long winters into chapter two, where they can arrive after the reader has met the woman in the cold house. Cut felt, very, just, and suddenly outright, and change she wondered whether selling would count to selling would count as mercy or theft stated flat. The images survive. The deductions disappear. That is the pattern for most failing pages: the material is fine and the order is wrong.
The opening line: short, placed, and doing one job
The 25-word limit exists because the first line has exactly one job, which is to place a person in a moment. Every word beyond that job is weight. Count the words in your opening line right now. If the count clears 25, the line is almost certainly doing two or three jobs: placing the person, explaining the past, and previewing the theme. Split the jobs across three sentences and the problem resolves.
Strong openers tend to follow a small set of shapes. A declaration of action: she counted the envelopes twice. A spoken line: which car, she said. A concrete image with a person inside it: the house kept the cold. A small disruption: the lid popped off and coffee ran across her wrist. Each shape is short because each carries one piece of freight.
Watch for three specific failure shapes. The weather-and-date opener spends twenty words on sky and calendar before any person appears. The biography opener introduces the character with age, occupation, and town, which is query-letter material, not first-line material. The philosophy opener states a general truth about life, love, or war, which asks the reader to admire the thought before caring about the thinker. All three can work in experienced hands. All three raise the odds of a fast skim on a debut.
A practical test: cover every line of your first page except the first, and ask whether the line alone would make a stranger read sentence two. If the answer is no, cut the line in half and keep the half containing the person. Repeat until the answer turns to yes.
Voice and action: giving the reader someone to follow
Readers attach to characters who speak and move. That is the entire basis of the voice-or-action check. A page that opens with a character doing something physical, saying something pointed, or thinking something sharp gives the reader a companion. A page that opens with fifteen lines of room description gives the reader a room.
Action does not mean combat or chases. Counting envelopes is action. Choosing the bench farthest from the gate is action. Standing too fast so a coffee lid pops off is action. The requirement is physical movement with a hint of intention behind it. Ask of your first paragraph: what is the character doing with their hands. If nothing, give them something to do.
Dialogue works even faster than action because speech carries relationship, class, region, education, and mood in a few words. One exchanged pair, such as which car and the gray one, go now, establishes two people, a plan, and urgency in eight words. You do not need a long conversation. One spoken line in the first 200 words changes the texture of the whole page.
Thought openings are the risky middle. Interior narration can carry a page when the thought itself is surprising or funny. Most thought openings on debut pages are summaries of feeling: she felt tired of everything, she had never belonged. Those lines ask for sympathy before earning interest. If your page opens in thought, test it by converting the first thought into speech or movement. When the conversion improves the page, keep the conversion.
Description openings work when the place is the event. A ferry terminal at dawn, a courtroom before sentencing, a kitchen where two people are not speaking: these carry tension because something is about to happen inside them. Pure landscape description without a person inside it almost never carries page one. Put the character in the place within two sentences, and give them a small friction with it.
Name density: counting what the reader must memorize
Every proper name on page one is a small debt. The reader meets Mara, Luis, Mercer, and Canal and must hold all four while also absorbing situation and tone. Two or three debts are fine. Eight or nine become fatigue, and fatigue on page one reads as rejection.
The tool counts capitalized words that do not begin a sentence, which approximates names while ignoring ordinary sentence starts. Fantasy and science fiction receive a higher limit because invented geography and titles are part of what those readers came for. A thriller naming six operatives, two agencies, and three cities on page one fails the same check a fantasy passes with six place names, and that difference is deliberate.
Audit your own names with a simple procedure. List every capitalized word on your first page that is not a sentence start. Circle the ones the reader must remember in chapter two. For every uncircled name, replace it with a role or description: the driver, the gray car, the riverside town. Introduce the circled names one at a time, each attached to a small action, so the reader meets the name and the person simultaneously.
First-name-only introductions usually beat full names on page one. Marisol Vega counted the envelopes is heavier than Marisol counted the envelopes when the surname does no work yet. Save surnames, titles, and place spellings for the moment each becomes necessary. Mystery writers should note one exception: a victim or client name that drives the plot can arrive early, because the story immediately starts using it.
Filtering, hedging, and tell words
Filtering words describe the perception instead of the thing: she felt the cold instead of the cold bit her wrist, he saw the door open instead of the door opened, she heard footsteps instead of footsteps climbed. Each filter adds a word and removes a sensation. Search your page for felt, saw, heard, noticed, realized, seemed, began to, started to, and wondered. Most occurrences can be deleted with the sentence improving.
Hedging words weaken the statement: very, really, just, quite, rather, somewhat, almost, suddenly. Suddenly deserves special attention because new writers use it as excitement punctuation. Excitement should come from the event, not the adverb. The lid popped off needs no suddenly. Delete the hedge and check whether the sentence still stands. It almost always stands straighter.
Tell-style words are a softer signal, and the tool treats them that way: 3 points each against a 20-point cap, rather than the heavier filter and backstory penalties. Was, were, been, told, explained, knew, and understood often mark summary where a scene could stand. She knew the town was corrupt summarizes. Show one corrupt exchange and the reader knows it without being told. That said, some was and were are structural and harmless. The tool counts them so you can see the total, not so you can reach zero. A page with two was forms and no other issues still passes comfortably.
Run the inventory in this order. First delete hedges, which are pure loss. Then unfilter perceptions, which usually shortens sentences. Then convert one summary passage into a brief scene. Stop when the counts sit under the limits. Chasing zero produces stiff prose, and stiff prose fails the read-aloud test that comes after every structural check.
Backstory: the history can wait one scene
Backstory on page one answers questions the reader has not asked yet. The valley where the mother grew up, the burned childhood home, the long winters of those days: each fact may be true and important, and each one pauses the present scene to deliver it. Agents cite early history dumps as a rejection signal because the dump suggests the writer does not trust the present scene to hold attention.
The repair is positional, not destructive. Nothing gets deleted from the novel. The history moves to chapter two or three, arriving at the moment a present-tense event gives it meaning. The burned house lands harder when the character stands in a cold house deciding whether to sell. The mother's valley lands harder when a letter arrives from that valley. Write the history on a separate sheet, finish page one in the present, then reinsert each fact where a scene earns it.
Two narrow exceptions exist. Memoir sometimes needs one orienting date or place early, because the contract with the reader includes real time. Even there, one clause beats one paragraph. And prologues built as history lessons usually confirm the agent's fear rather than calming it. If your opening requires three paragraphs of prior events to make sense, the opening starts too late or too early. Move the start line until the scene explains itself.
Test your page with a highlighter. Mark every sentence that describes events before today. If marked sentences appear in the first 200 words, move them. When zero remain, the backstory check passes and, more importantly, the page starts moving at full speed from the first line.
Dream, wake-up, and thought openings
Openings where the character wakes up, gets out of bed, showers, dresses, eats breakfast, and commutes are sometimes called morning-routine openings. They fail for a structural reason: routine contains no decision and no friction, so nothing distinguishes this morning from any other. If your story starts with a morning, start with the morning that differs: the phone already ringing, the door already open, the envelope already missing.
Dream openings fail harder because they spend the reader's first trust on events the story then cancels. The reader invests in the chase, the kiss, or the disaster, learns it did not happen, and must start investing a second time on page two. Some readers forgive this once. Tired professionals reading submission 200 of the week rarely do.
Thought openings sit between the two. A sharp, specific thought can open a novel brilliantly. The failure mode is the general thought: reflections on belonging, tiredness, hometown weather, or the strangeness of returning. Convert general thoughts into particulars. Instead of she had never belonged here, try she still reached for the light switch on the wrong side of the door. The second version is a thought wearing a body, and bodies hold readers.
If your draft opens with any of these three, try the cut test. Delete the first two paragraphs and read from paragraph three. When the page reads better, the opening was throat-clearing and the cut is the fix. When the page reads worse, restore the paragraphs and sharpen them instead. Either outcome takes five minutes and answers the question with evidence instead of worry.
Genre-specific guidance
Thriller and suspense pages live or die on motion. Open with a person doing something irreversible or time-pressured: counting, carrying, running, calling, deciding. Keep names to two or three. Favor short paragraphs. End page one mid-motion rather than on reflection, because the promise of a thriller is what happens next. Filtering words hurt thrillers more than any other genre, since distance kills urgency. Cut every felt and suddenly first.
Mystery and crime pages need a body, a question, or a wrongness within the first 300 words. The detective can arrive ordinary, but something in the scene must be off: the lid that should not have popped, the kiosk man watching too long, the coffee the character did not want. Introduce the victim or client name early if the plot uses it immediately. Hold back the detective's personal history for chapter two. Backstory about divorces and dead partners is the classic mystery page-one failure.
Romance pages need two things fast: a voice the reader wants to spend 300 pages with, and the charge between people. Dialogue is the highest-value tool here. One exchanged pair with wit or friction outperforms three paragraphs of attraction described from inside one head. Physical description of the love interest can wait. The feel of being near them cannot. Avoid opening with the heroine reflecting on past relationships. Start with the present collision.
Fantasy pages carry the heaviest naming load, and the tool's higher limit reflects that. Spend the budget on names that do immediate work. A city where the scene happens earns its name. A dead king from three centuries back does not. Open with a character want inside the strange world rather than with world mechanics. Magic systems, currencies, and histories belong after the reader cares about the person carrying them. One sensory detail of strangeness beats one paragraph of explanation.
Science fiction pages face the same budget problem with technology instead of geography. Open with the human situation the technology complicates, not with the technology. A woman smuggling envelopes matters before the reader knows how the surveillance grid works. Explain exactly enough for the scene to function, and mark every unexplained term to pay off within five pages. Readers tolerate mystery. They do not tolerate lectures.
Literary fiction pages get the most freedom and face the subtlest test. Voice can carry an opening with almost no plot, but only when the voice is specific enough to be unmistakable. Generic lyricism about light, memory, and houses fails the voice check in spirit even when it passes the letter of it. The test for a literary opening: could this paragraph appear in any novel. If yes, replace abstractions with particulars until the answer is no. The house kept the cold the way some people keep grudges works because grudges make the cold a character trait.
Memoir pages need orientation plus intimacy. One clause of when-and-where earns the reader's bearings, then the page must deliver a lived moment with sensory detail. Memoir fails most often through summary: years compressed into a paragraph of and-then. Slow down. One scene, fully inhabited, outperforms a decade summarized. The backstory check still applies, because even true history stalls a present scene when it arrives too early. Start inside the memory, not above it.
Edge cases and failure modes of this tool
The tool reads only what you paste, up to 4000 characters. Paste 60 words and the name limit shrinks with the word count, which can make a sparse fragment look cleaner than it is. Paste the densest 4000 characters of chapter one instead of the actual opening and the score describes the wrong page. For honest results, paste the first 200 to 300 words exactly as they appear in the manuscript, line breaks included.
The capitalized-word count is an approximation. It skips sentence-initial words to avoid counting ordinary vocabulary, but it still counts capitalized words that are not names, such as Monday, Christmas, or Spanish. A page heavy with such words can fail the name check without naming anyone. When that happens, read the flagged count with judgment: if the capitalized words are calendar terms rather than characters, the page is fine and the check is noisy. Rename nothing.
Dialogue detection looks for quotation marks. Pages using em-dash dialogue conventions or indirect speech without quotes will read as speechless to the tool. If your manuscript uses dashes for speech, the voice check may fail on a page full of voices. Tell the opening-type selector the truth, add one quoted line for the test if needed, and interpret that check lightly.
Very short prose, flash openings, and experimental pages can fail the line check in misleading ways. A one-line paragraph of 30 carefully built words fails the 25-word limit while succeeding as art. The limit encodes what tired professionals cite as a skim signal, not what literature allows. Break the limit deliberately when you understand the cost, and keep the receipt: a 30-word opener should place person, moment, and tension all at once.
The score cannot see concept, comps, query strength, or agent fit. A page can score 95 and still collect rejections behind a quiet premise or a mismatched batch of agents. It can score 62 and still win a request from an agent who loves the exact situation. Treat the tool as a friction meter for craft, not a prediction of any reader's taste. Its honest scope is narrow: it assembles counts from your pasted text with fixed rules and states plainly where commonly cited turn-offs appear.
The tool also cannot see revision three. Writers sometimes run the check after every paragraph and sand the page smooth until no check fails and no life remains. That is misuse. Run it on drafts, fix the top miss, and return to writing. A page that passes all six checks but bores its writer aloud needs new material, not new counts.
Questions writers ask next
How many names are too many on page one. For most genres, two named people plus one named place is comfortable. Three named people works when each one acts in the scene. Beyond that, every additional name should be cut or delayed. The tool's limit scales with pasted length, so a 300-word page allows a few more than a 150-word fragment, but the reader's memory does not scale. When in doubt, keep two.
Should the opening line name the character. Naming in line one is optional. What matters is that a person is present, named or not. She counted the envelopes works. Marisol counted the envelopes works slightly better when the name sounds like the book. Never delay the person past line three for the sake of atmosphere.
Can description open a novel. Yes, when the place contains tension and a person enters within two sentences. The ferry terminal passage in the sample default works because the woman acts inside it from line one. Description of an empty landscape for a full page works only for writers whose track record buys patience. Debut pages buy no patience. Spend description on friction between character and place.
Is starting with dialogue acceptable. Agents request pages that open with speech every day. The caution is that unattributed speech from strangers can confuse, so anchor the first spoken line fast: which car, she said, names speaker and situation in four words. Avoid opening mid-argument without orientation, since the reader cannot tell the stakes of a fight between unknowns. One calm, pointed line beats one loud, context-free line.
How much backstory is safe on page one. Zero flagged phrases is the tool's line, and it matches what most agents advise debuts. One short clause of orientation, such as after the funeral or since the layoff, rarely hurts when the scene immediately moves. A paragraph of history always hurts. The difference is whether the past explains the present action or replaces it.
What if the story requires a dream. Write the dream, then relocate it. Dreams work as chapter-two memories, as recurring images, or as prologues in experienced hands. They fail as page-one events because the reader's first investment deserves real stakes. Keep the best image from the dream and spend it where it cannot be accused of stalling the plot.
Does the tool store or share my text. No. The checks run deterministically on the pasted words inside the page, and the counts describe only that pasted text. Nothing is uploaded, compared against any database, or used to train anything. The honesty of the tool is structural: fixed word lists, fixed limits, fixed deductions, visible on this page.
How does this check relate to line-level editing. It does not replace it. This audit screens for six structural turn-offs. Repetition, rhythm, dialogue tags, and filter words across a full chapter need different instruments. Run this diagnostic first to fix page-one structure, then audit the chapter with the dialogue tag and filter word tools before querying.
Turning the diagnosis into a revision
Work the recommendations from the top. The tool orders fixes by cost, so the first item is the change that buys the most score and the most reader patience. Fix one item, repaste the page, and run the check again. Single-issue pages routinely move into the strong band with one edit, and watching the score move confirms the edit worked.
Revise in this sequence. First, fix the opening mode: cut dreams and morning routines so the page starts in the real scene. Second, shorten the opening line to under 25 words. Third, move backstory off the page. Fourth, delete hedges and unfilter perceptions. Fifth, trim names toward the limit. Sixth, add one spoken line or physical action if the page still shows none. Each step is small, each is verifiable by rerunning the check, and the page compounds from failing to strong without a full rewrite.
Then read the page aloud. The tool cannot hear rhythm, and rhythm is what carries a structurally clean page from correct to compelling. Mark every sentence where your voice stumbles and rewrite those sentences until they speak smoothly. Follow with one promise test: ask what the final line of the page makes the reader want. If the answer is nothing specific, sharpen the last line until it points at trouble, a question, or a decision.
Do the revision work where the draft lives. Co-Writer at /write is the place to carry out the fixes this tool plans, because the manuscript can be rewritten, rescored in spirit, and expanded into chapter two in the same session. Paste the revised opening back into the checker above after each editing round to confirm the misses clear. When the page holds a strong score and reads smooth aloud, draft the next chapter there too, so the momentum of a fixed page one carries straight into the pages that follow.
A final note on rules and judgment
Every limit on this page is a stand-in for a tired professional's first impression. Twenty-five words is not a law of literature. Three filtering words is not a ceiling on style. Zero backstory flags is not a theory of narrative. They are tripwires that experienced readers report tripping over, encoded as counts so you can see your own page through those readers' eyes for thirty seconds.
Break any of them on purpose when the story earns it. A 30-word opener that places person, moment, and tension simultaneously earns its length. A needed place name earns its memory debt. A memory arriving on page one earns its slot when the present scene cannot function without it. The tool's job is to make the cost visible, not to forbid the purchase.
What the tool forbids is accident. No page should fail these checks because its writer never counted the line, never noticed the dream frame, or never saw the history dump sitting on top of the scene. Run the check, see the counts, make the choices with open eyes, and send pages that represent the book at its best. That is all any diagnostic can do, and for page one it is enough.