Competition-grade screenplay PDFs · Remote MCP service
Competition-grade screenplay PDFs, from any AI.
Narracore Screenplay Formatter is a remote MCP tool. Paste rough script text into Claude, ChatGPT, Cursor, or Gemini — get back a competition-grade, submission-ready screenplay PDF on the same fixed typewriter grid that powers narracore.cn: one page ≈ one minute, US Letter or A4, US/UK/French submission styles, (MORE)/(CONT'D) dialogue breaks, and an anonymous title page for blind review. It also speaks film: give your agent the film-breakdown protocol, and its own shot-by-shot observations become validated, comparable, compiled results — without the video ever leaving your machine.
How it works
Three steps, one sentence.
01
Connect once
Add the MCP URL to your AI app as a custom connector. No account, no API key to start.
02
Say what you want
“Format this screenplay.” Paste or attach your rough script text — Chinese, English, or mixed.
03
Get the package
Structured blocks, a clean plain-text screenplay, and a download link to a competition-grade, print-ready PDF.
Competition-grade PDF
Built to submission standards, not to a printer's convenience.
01
One page ≈ one minute
Every character sits on a fixed typewriter grid — 61 × 55 per Latin page, 33 × 38 per CJK page. The page count is computed by the layout engine, so timing estimates hold on every viewer and printer.
02
US, UK & French submission styles
US Letter or A4 per style: BBC-style margins and double-sided scene numbers for the UK, a centered French dialogue column for France. One option — style="uk" — no re-typing.
03
Dialogue breaks & blind review
dialogue_breaks=true adds (MORE) at the page bottom and SPEAKER (CONT'D) at the next page top wherever dialogue spans a break. cover.anonymous=true prints a title page without the author — the blind-review convention.
04
Vertical Japanese manuscript sheets
Three sheet formats typeset vertically (columns right to left): 業務用 20×10, the 400-character 20×20, and a Kido-style 20×40 on A4 landscape — with kinsoku line-breaking and upright tate-chū-yoko digits.
05
Genkō yōshi sheet counts
analyze_screenplay reports your script's sheet count under three competition conventions (400-char 20×20, 200-char 20×10, Kido-style 20×40) in one pass. Always confirm the current rules of the competition you are entering.
06
Bilingual side-by-side PDFs
Bring your own translation: render_bilingual_pdf typesets the original and the translation side by side on A3 landscape, aligned scene by scene. The same original stays free to re-render for 24 hours.
We follow the format standards these submissions expect — fixed grids, margins, scene-number and title-page conventions. No competition has endorsed this service; always check each competition's current rules.
Before / after
Rough text in. Professional screenplay out.
You paste
Your AI answers with
Scene headings are numbered automatically, dialogue is indented to the dialogue column, transitions align right, and the whole document paginates on a fixed grid — the same algorithm as the narracore.cn editor.
From page to screen
Your screenplay, turned into video-prompt scaffolds.
01
Structured scaffolds
build_video_prompts splits each scene into ~15 second segments and shots, registers every character, and pre-fills dialogue lines verbatim from your script. The creative fields stay yours to fill.
02
Drift-proof checks
check_video_prompts (free) checks your finished prompts against the screenplay: dialogue must match verbatim, cut points must carry over shot to shot, and shot blocks never restate shared settings.
03
Style is yours
A guided decision checklist, three editable presets (engine cinematic / hand-drawn on-twos / live-action), or your own style note passed through verbatim. We never decide your look for you.
Your screenplay
Scaffold your AI fills in
Deterministic structure and verbatim dialogue anchors — your AI fills performance, composition and camera; segment durations are estimates you can adjust. We format and verify; we never generate video.
Film breakdown · for AI agents
Your agent watches the film. We make it computable.
01
A shared film language
Free — get_film_contract hands your agent:
- Closed vocabularies: shot size, angle, movement, blocking, dialogue beat, performance, sound
- An alias table, so “CU” and “close-up” mean the same thing
- Observation and reasoning prompt templates — interpretive fields stay marked as interpretations
- Long-take window protocol and hard limits
02
You observe, we verify
Free — your agent watches the film locally, with any vision model:
plan_film_analysis— cut timestamps → the analysis windows everyone sharesvalidate_film_breakdown— vocabulary, cross-references, long-take window completeness- Only structured notes reach the server — the video is never uploaded, and nothing is stored
03
Compiled into something you can query
One paid call — compile_film_breakdown — turns notes into results:
- Long-take windows merged back into single shots, so a 96-second take counts as one shot
- Exact statistics: shot sizes, pacing, shot lengths, per-scene aggregates
- Cross-shot performance arcs and cut logic
- A PDF report — scene table, arcs, shot tables
Your agent’s notes · one window per long take
Compiled · what you get back
Honest by construction: interpretive fields are marked as interpretation, heuristic reconstructions are labelled as reconstructions, and a long take with missing windows is rejected instead of guessed. Contract, planning and validation are free; compiling is billed per film analysis — recompiling the same analysis within 24 hours is not charged again.
Connect your AI
One URL, any MCP client.
https://mcp.narracore.app/mcp
copied
Claude (claude.ai / Desktop)
- Open Settings → Connectors
- Add a custom connector, paste the URL above
- Then paste the starter prompt below
ChatGPT (Developer mode)
- Enable Settings → Apps & Connectors → Developer mode
- Create a dev connector with the URL above
- Then paste the starter prompt below
Cursor
- Settings → MCP → Add new global MCP server
- Type:
url, value: the URL above
Anything else that speaks MCP
- Streamable HTTP endpoint, no auth header needed
- Tools (15):
format_screenplay·analyze_screenplay·diagnose_scenes·convert_fountain·render_bilingual_pdf·build_video_prompts·check_video_prompts·get_film_contract·plan_film_analysis·validate_film_breakdown·compile_film_breakdown·prepare_story_window·absorb_story_window·render_story_handoff·check_credits - Works with Gemini API MCP support and every MCP SDK client
Starter prompt — copy it, paste it into the chat
You are now connected to the Narracore Screenplay Formatter MCP service (https://narracore.app). Follow this whenever I ask about screenplays or scripts:
1. Confirm the connection in one line, then ask me to paste a script.
Do not try to verify or configure anything in a browser.
2. When I paste text (Chinese, English, or mixed), call format_screenplay on it.
Give me the PDF link (valid 24h) and page count. Ask whether a submission
style is needed (us/uk/fr, Japanese vertical sheets, dialogue_breaks=true,
anonymous cover). If validation_status is not "ok", also show the
diagnostics and ask me before changing anything.
3. Free tier: 1 free credit per day, no key needed. If a tool responds with
free_tier_exhausted or quota_exhausted, relay the price and purchase link
from that response verbatim, then wait for me to buy and paste back a
license key (format NRC-XXXX-XXXX-XXXX-XXXX). Never invent a license key;
never guess a price.
4. When relevant, offer the other capabilities: script analysis and diagnosis,
video-prompt scaffolds, film-breakdown compilation, long-story continuity.
The checker and contract tools cost nothing.
5. My text is processed in memory and never stored; PDF links expire after
24 hours. Full service manual: https://narracore.app/llms.txt
Reply in my language.Paste it into the same chat right after you connect. It is written in English on purpose — your AI will read it and reply in your language.
Pricing
One free credit, every day.
Free tier
$0 / daily
- One free credit every day — no account, no key (one screenplay for 24h, one film analysis, or one story project)
- Extra credits come in packs — no subscription
- Paid packs are live: $1.49/10 credits, $5.99/50 credits
Starter
$1.49 / 10 credits
- 1 credit = one screenplay for 24 hours across every screenplay tool — format, analyze, diagnose, convert, video prompts
- License key works in any AI app
- Failed calls are never charged
Studio
$5.99 / 50 credits
- ≈ ¥40 · best value per credit
- Same key, any app, no subscription
- Refund within 7 days if unused
Payments are open (PayPal, cards inside the PayPal window). The free daily quota keeps working without a key. Pricing in USD; refunds within 7 days if unused; failures are never charged. No subscriptions: you buy credits and use them — one credit covers one screenplay for 24 hours across every screenplay tool (format, analyze, diagnose, convert, video prompts); one credit covers one film analysis (same analysis recompiled within 24 hours is not charged again); one credit covers one story project for 24 hours. Contract, planning and validation for film breakdowns — and story preparation and handoff — are always free.
FAQ
Questions, answered.
What is MCP?
Model Context Protocol — an open standard that lets AI apps like Claude, ChatGPT, Cursor, and Gemini call external tools. You add our URL once; your AI can then use screenplay formatting whenever it's relevant.
Do you store my screenplay?
No. Your text is processed in memory to produce the result. The generated PDF sits behind an unguessable signed link and is deleted after 24 hours. We don't log script content and never use it for training.
Which languages are supported?
Chinese, English, and mixed text. The layout engine picks the right track automatically: a 33-char × 38-line CJK grid or a 61 × 55 Latin grid, both on US Letter paper.
Why a fixed character grid?
It's how professional screenplay paper works: every character occupies one cell, so one page always equals roughly one minute of screen time. Your PDF will match narracore.cn's print preview line for line.
What if the structure detection is unsure?
Every result carries a confidence score and line-level warnings (e.g. “line 12 might be dialogue”). Your AI will offer to fix ambiguous lines and re-run — re-formatting is a normal call, and failed calls are never charged.
Can it analyze a film — shot breakdowns, pacing, performance arcs?
Yes — as a protocol for your AI, not as a black box. Your agent watches the footage locally (any vision model), writes shot notes against our film contract (closed vocabularies for shot size, camera, blocking, performance, sound), and this service plans the analysis windows, validates the notes, merges long takes back into single shots, and compiles exact statistics plus cross-shot performance and cut logic into a PDF report. Free tools: contract, planning, validation. Compiling is billed per film analysis; the same analysis within 24 hours is not charged again.
Do you upload or store my video?
No. The original video is never uploaded: it stays on your machine, where your agent watches it. Structured breakdown input exists only while a request is being processed and is not persisted as database records. PDF reports you generate are stored temporarily and deleted after 24 hours.
Refund policy?
Full refund within 7 days if you haven't used any formats. Service failures are never charged in the first place. Contact huiyicd@foxmail.com.
Is the PDF really competition-grade?
The fixed 55-line grid is the one-page-per-minute standard that screenplay format is judged against — on US Letter or A4, with US/UK/French style variants, (MORE)/(CONT'D) dialogue breaks, and an optional anonymous title page for blind review. These are format standards, not an endorsement by any competition; always check each competition's current rules before submitting.
Does it support Japanese vertical manuscripts or bilingual drafts?
Yes. Three vertical Japanese sheet formats (20×10 business, 20×20 400-character, 20×40 Kido-style) with kinsoku line-breaking, plus sheet counts under all three conventions. For submissions that need a translation, render_bilingual_pdf pairs your original with your own translation side by side, scene by scene — it typesets translations, it never translates for you.