ScriptForge AI - Create scripts with appropriate image generation prompts

ScriptForge AI — ready-made scripts for your channel

ScriptForge AI studies the channels you indicate, understands their subject and writing style, creates video ideas on its own and writes the entire script. Afterwards, he does the narration, marks the second in which each sentence is said and writes image requests for each scene. Everything within one program, on your computer.

Program for Windows. You download, unzip and double click. You don't need to install anything else. When first opened, it asks for the serial number of your license Like Points and releases the artificial intelligence of your plan.

Don't have a license yet? Click here to get your serial number.

What's new in this version

• The script does what the title promises. If the title says "3 miracles" or "5 prophecies", the text delivers the three, the five — each one announced clearly and truly developed, no messing around and just delivering the first one. At the end, the program rereads the script, checks item by item and rewrites what was missing before showing it to you.
• Download all scripts at once, already organized in folders. One click generates a compressed file with a numbered folder for each script. No more dragging thirty files by hand.
• Ready publishing package. Each script comes with three title options for you to test which one yields the most views, the description already broken into lines and the hashtags. All of this is already included in the downloaded file.
• Report problem button. Did it make an error? One click downloads a text report with everything that happened, ready to send to the support WhatsApp. It goes off on its own after 2 days.
• Choosing writing quality for those who registered their own artificial intelligence key on the panel.
• Clearer and more readable screen, with adjusted contrast.

Step by step — from scratch to finished script

1. Open the program and enter the serial number. It verifies your account and releases the artificial intelligence of your plan. You only need to do this once.
2. Create a project (Projects tab → Create project). Give it a name, say the main subject, choose the language, narrative style and average length of your videos. Paste the links of the channels you want to use as references, one per line.
3. Click "Save and study channels". The program reads the most viewed titles and videos on each channel and creates a picture of what works there: subject, way of writing, title format.
4. Go to "New script" and click "Generate ideas now". He creates the ideas alone, in the style of the studied channel. Choose how many you want: 5, 10, 20, 30 or 50.
5. In the Titles tab, select the ones you want and click on "Generate itineraries for those marked". You can mark several at once by holding Shift, or use "Mark all".
6. Follow the Production tab. The scripts are written one at a time, in parts, so as not to fail in the middle. You see the progress in percentage.
7. Adjust text in the Editor if you want, and download it in TXT, DOCX or PDF — or download them all at once in separate folders.
8. Generate the narration in the Narration tab, mark the speaking times e create the illustration prompts if you are going to use images.

What each screen does

Projects

Each project stores a reference channel (or several), the language, style, length of the videos and the folder where everything is saved. It allows you to play more than one channel without mixing things up.

• Referral channels — paste one link per line. It could be the link to the channel or a video of it.
• Study channels now — redo the study when the channel changes direction. The result appears as a list of subjects that support that channel, and you can open the complete diagnosis.
• Folder where to save — optional. In white, everything is inside the program folder, organized by project.

New script

This is where ideas are born. The main button already does everything on its own with what has been learned from the channels. If you want to control something, open "Optional settings":

• Request from you for these ideas — blank, the program decides on its own. Satisfied, he obeys ("in this group, prefer stories about common people").
• Learn from more material — paste the link to a channel or specific videos, send text files or paste a script you like. He learns the rhythm without copying any phrase.
• Which videos to use — most viewed (default), newest, oldest, or random. And how many to analyze: from 1 to 15.
• Language — 22 languages available, from Portuguese to Japanese.
• Words per script — calculated alone by the duration you requested. Only move if you want an exact size.
• Rhythm of the narration — how many words per minute the narrator speaks. It is already adjusted by language (150 in Portuguese).
• What CANNOT appear — themes, words and tones that you don't want in the text.
• Call to action at the end — the invitation that closes the video (comment, subscribe, share an experience).
• Quality of writing — only appears for those who have registered their own artificial intelligence key on the website panel. See the keys section below.

Available narrative styles: Documentary, Biblical History, Dark Channel, Mystery, History, Curiosities, Technology, Investigation, True Crime, Psychology, Storytelling, Motivational, Finance, Science, Cinema, Biography, Review, Top 10, Ranking, Tutorial, Podcast, Video essay, Mini documentary, Cinematic narrative, Netflix Style, MrBeast Style, SunnyV2 Style and MagnatesMedia Style.

Titles

List of suggested ideas. Mark the ones that will become a script. It has search by word, filter by batch, mark/unmark/reverse everything, favorite and delete in batch. The duration chosen here applies to all itineraries in that batch.

Production

It shows the progress of each script being written and the list of those that have been completed. This is where the batch download:

• Choose the format: TXT, DOCX, PDF or Markdown.
• Leave marked "A folder for each script" and you receive a compressed file with folders already numbered — 01, 02, 03 — each with the name of the script. Uncheck if you prefer all loose files together.

Editor

Write, delete and add as you wish. Real-time word and video minute counter. Buttons to download in TXT, DOCX or PDF, open the script folder and go straight to the narration.

• Magic wand — select an excerpt and choose what to do with it: rewrite it with other words, make it shorter, develop it better, explain it simply, increase the tension or correct the Portuguese. You can add your own tip ("cite the year and the name of the ship, and end with a question"). Only the selected section changes, the rest remains intact.
• To publish on YouTube — three alternative titles, the description in lines and the hashtags. If an old script doesn't already have this block, the "Assemble titles and description" button creates it right away.

Narration

Transform the script into audio. The file is saved in the script's own folder, with the same name as the text.

• Voice language and voice — hundreds of voices, male and female, in more than 140 languages and accents.
• Speed — from -50% (slowest) to +100% (fastest). Default: 0.
• Tom — makes the voice lower or higher. Default: 0.
• Listen to a preview — tests your voice on a short phrase before spending time on the entire script.
• Cloned voice (optional) — send 10 to 30 seconds with the voice you want to copy and the program will narrate with it. You need to install cloning once, using the button on the screen; the program does the installation itself and shows the progress.

Speech times

From the generated audio, the program marks the second each sentence is said and saves the ready subtitle. It is used to synchronize images and cuts when editing the video. The script must already have audio.

Illustration

Write a detailed image request for each spoken part, keeping the same characters and the same setting from the beginning to the end of the video — this is what prevents the protagonist from changing faces in the middle.

• Visual styles: Realistic/Cinematic, Anime, 2D Cartoon, 3D Pixar/Disney, HQ/Comic book, Watercolor/Painting, Dark/Suspense, Documentary/Journalistic, Cyberpunk/Futuristic, Minimalist/Infographic, Clay/Stop motion, Retro/80s VHS and Personalized (you write).
• Video format: lying (16:9), standing (9:16) or square (1:1).
• Extra Direction: a detail that applies to all scenes ("always at night, with rain").

The itinerary must already have the times marked.

Style Library

Send your texts in a text file and the intelligence learns your voice, your rhythm and your personality — without copying any sentences. Here is also the list of expressions that should be avoided: write the clichés you hate ("embark on this journey") and they no longer appear in the scripts.

Help

• Download the report (TXT) — the file with everything the program did. Contact support at WhatsApp when something goes wrong. It saves the last 2 days and turns itself off after that.
• Copy the report — to paste directly into the chat.
• Save the report to the Desktop — leaves the file ready there, without you looking for any folder.
• Program status — shows what’s working and what’s missing. The "Recheck my account" button updates your license and intelligence without closing the program.

Where the files are

Each script gets its own folder, within the project folder. Text, audio, speaking times, and image prompts from that script all live together there. The "Open folder" button in the Editor and on the Times screen takes you directly to it. If you don't choose a folder when creating the project, everything is organized within the program itself.

Required API Keys

API key is the password that authorizes the program to use artificial intelligence. You do not enter any key within the program — it searches for the ones in your plan on its own as soon as you enter the serial number. Registering your own on the panel is optional and helps you not depend on the shared quota, in addition to freeing you to choose the quality of the writing.

• OpenAI (recommended) — writes the scripts and serves as a reinforcement in marking the times of speech. Register on the panel · Watch the video tutorial
• Google Gemini (optional) — write alternative when the first one fails. Register on the panel · Watch the video tutorial
• OpenRouter (optional) — another writing alternative. Register on the panel · Watch the video tutorial
• Anthropic (Claude) (optional) — another writing alternative. Register on the panel · no video
• AssemblyAI (optional) — marks the second in which each sentence is said, on the Speech times screen. Without it, the program uses other available paths. Register on the panel · Watch the video tutorial

Writing quality: Whoever registered their own OpenAI key on the panel gains, on the New script screen, the option to choose between fast and economical writing (default), more careful text, long texts with more consistency or a balance between the two. Those who use the shared quota do not see this option — in this case, the configuration defined by the administrator applies, and consumption is divided among all students.

Requirements and limits

• Windows. Nothing more to install: the program is complete.
• Need internet. The license is granted with each opening and the artificial intelligence runs in the cloud.
• The license is personal. It identifies your account and releases what your plan offers.
• Accepted video length: 1 to 180 minutes. Very long scripts take longer to complete.
• Library example texts must be text (.txt) files.
• Sample to clone voice: 10 to 30 seconds, with clean audio and just one person speaking.
• Voice cloning is installed only once and takes up space on your computer; on a weak machine it takes longer to narrate.

Common problems and how to solve them

• "Your license needs to be verified to continue" — your subscription has expired, been cancelled, or the internet has gone down. Check the connection and click "Recheck my account" on the Help screen. If it persists, renew your license here.
• No new ideas appear — the project has not yet studied any channels. Go to Projects, paste the links and click "Study channels now".
• "Choose a voice before generating audio" — first select the voice language and then the voice from the list on the side. Use "Hear a preview" to check it out.
• Audio is not generated — check if you chose a script from the selector at the top and a voice. If the message says a file is missing from the package, the program has been decompressed halfway: delete the folder and unzip the downloaded file again, in its entirety.
• "Choose a script that already has audio" — speech times come out of the audio. Generate the narration first.
• "Choose a route that already has the times marked" — image prompts move with the times. Go through the Speech Times screen first.
• The script came out shorter than you asked for — increase the duration in minutes, or the number of words in optional settings.
• Old scripts without publication titles and description — open the script in the Editor and click on "Add titles and description".
• Any other error — go to Help, click on "Download report (TXT)" and send the file to support at WhatsApp. It says exactly what happened, and no password or key of yours appears on it.