Free AI video prompt library
135 copy-ready prompts, format recipes and Claude Code skills for making video with AI, each marked with the models it works with. Free, no account.
Most copied so far
A fill-in template that gives any video model the seven things it needs: subject, action, setting, camera, light, style and sound. Fewer vague words, fewer wasted generations.
A filled 7-part prompt for a hand-drawn anime shot: flat cel shading, animation on twos, a held painted background, and the three things the style breaks in current video models.
A character description block you reuse verbatim in every prompt, plus a generated reference sheet, so a face, outfit and build stay the same from shot to shot.
The 14 camera moves that generation models respond to reliably, with the phrasing that works and the phrasing that produces a wobbling mess.
One agent prompt that reads your transcript and writes ten generation prompts for cutaways that show what is being said, each tagged with the sentence it covers.
The before/after format with the twist that keeps it from being skipped: open on the finished result, then show how bad it was, then the change. The prompt returns a timed script and text plan.
Filter by category, tool, platform, level
135 entries
Entries
Before you post a generated video: disclosure and rights checklist
Six checks for AI-generated clips before they go on TikTok, Reels, Shorts or a client's site: labels, likeness, music, logos, model terms and provenance.
Anime style: 2D cel look, limited animation, clean line weight
A filled 7-part prompt for a hand-drawn anime shot: flat cel shading, animation on twos, a held painted background, and the three things the style breaks in current video models.
Architectural walkthrough: an interior on a steadicam in natural light
A filled 7-part prompt for a real-estate or design walkthrough: one slow steadicam path through a room, windows as the only light, straight verticals, and the turns that make geometry drift.
Generate a 10-clip b-roll pack for a talking-head video
One agent prompt that reads your transcript and writes ten generation prompts for cutaways that show what is being said, each tagged with the sentence it covers.
Camera movement vocabulary that video models understand
The 14 camera moves that generation models respond to reliably, with the phrasing that works and the phrasing that produces a wobbling mess.
Keep the same character across clips: the reference sheet prompt
A character description block you reuse verbatim in every prompt, plus a generated reference sheet, so a face, outfit and build stay the same from shot to shot.
Establishing shot: open a video with a place
A prompt for the wide opening shot that tells the viewer where they are, with the three things that make a generated landscape look filmed rather than rendered.
Claymation and stop motion: 12 fps feel, fingerprints, slight jitter
A filled 7-part prompt for a clay stop-motion shot where the imperfections are the point: visible fingerprints, a frame-by-frame stutter, a set that looks built by hand, and what not to ask for.
Fashion lookbook: a model walks in a studio at 24 fps
A filled 7-part prompt for a lookbook clip: seamless backdrop, one walk toward the camera, cinema frame rate so the fabric moves like fabric, and the faces, feet and prints that need extra care.
Fewer warped hands and faces: what to change in the prompt
The prompt-side fixes for the classic generation failures (melting hands, drifting faces, extra limbs), ordered by how often each one works.
Food close-up: macro, steam, a slow pour
A filled 7-part prompt for a food shot that looks edible: one macro framing, one slow physical event (a pour, steam, a crack), side light that shows texture, and the hands and labels to keep out.
Ask an agent to turn a script into a numbered generation shot list
A prompt for Claude or ChatGPT that converts a 30-second script into 6-8 shot prompts in the 7-part structure, ready to paste into a video model one by one.
Image-to-video: generate the first frame, then animate it
Make a still with an image model first, fix the composition there, and only then ask the video model to animate it. Cheaper iterations and the look you chose, not the one the video model guessed.
Music visualizer: abstract particles that pulse to a stated tempo, loopable
A filled 7-part prompt for an abstract background clip: particles that pulse at a stated BPM, a near-seamless loop, no text; the model cannot hear your track, so the beat is matched in the editor.
Nature documentary: long lens, slow motion, no narration
A filled 7-part prompt for a wildlife shot that looks filmed from a hide: 600mm compression, heat haze, one animal doing one slow thing, slow motion at a stated rate, and the anatomy traps.
Pixel art: an 8-bit game cutscene on a locked pixel grid
A filled 7-part prompt for a retro game cutscene: a stated sprite size, a palette of 16 colours, parallax layers, no sub-pixel smoothing, and why a still from an image model is the safer first frame.
Product hero shot: a 5-second generated clip for an ad or a listing
Prompt for a clean product shot with one motion and one light change, the kind that sits at the top of a listing or opens a 15-second ad.
Retro VHS: a 1990s home video shot on a camcorder
A filled 7-part prompt for the camcorder look: 4:3 inside the frame, tracking noise, colour bleed, auto-focus hunting and a zoom nobody meant, with the date stamp added in the editor.
Style lock: one paragraph that makes every clip look like the same film
A reusable style paragraph (stock, palette, grain, light rules) appended to every generation prompt so a series cuts together instead of looking like ten different ads.
Talking-head clip: the script format AI avatar and lip-sync tools read best
How to write a 20-second spoken script so that a lip-sync or avatar model delivers it with natural pauses and no mispronounced numbers.
UGC-style generated ad: make it look like a phone, not a studio
Prompt rules that push a video model toward the handheld, imperfect look of user-generated content.
The 7-part prompt structure for text-to-video
A fill-in template that gives any video model the seven things it needs: subject, action, setting, camera, light, style and sound. Fewer vague words, fewer wasted generations.
Watercolor and painterly: a moving painting with paper grain and bleeding edges
A filled 7-part prompt for a hand-painted look: wet-on-wet edges, visible paper grain, boiling pencil lines, slow motion so the paint stays paint, and the drift to 3D when the action is too fast.
Before/after: the 20-second reveal with the 'after' shown first
The before/after format with the twist that keeps it from being skipped: open on the finished result, then show how bad it was, then the change. The prompt returns a timed script and text plan.
Build in public: a 45-second behind-the-scenes cut around one decision
The founder diary format that earns trust because it shows the decision, not the launch: the prompt structures a week of real screen and desk footage around one choice and one measured result.
Day in the life: a 45-second cut from real footage with one through-line
A recipe for the vlog-style day that keeps working because the viewer waits for one promised payoff; the prompt turns your clip list into a timed script with text overlays.
Founder explainer: 'I built this because' in 40 seconds, to camera, one problem
The founder-to-camera format that works because it names the specific moment the problem became unbearable, then shows the product once. The prompt bans origin-story clichés and budgets the words.
'I tried it for 7 days': a diary cut with one metric logged every day
The experiment-diary format that stays honest because one number is measured daily and shown on screen; the prompt plans the daily clip, the tally and a verdict that includes what did not work.
Countdown listicle: 5 items in 45 seconds, best one last, numbers on screen
The ranked list format, with the rule that makes it hold: items 5 to 2 get equal time and number one gets the longest shot. The prompt returns script, text plan and the caption list.
Myth vs fact: three beliefs checked in 35 seconds, each with a demonstration
The format where a common belief is stated, then tested on camera, then labelled. Works because the viewer picks a side in the first second; the prompt insists on a visible test per myth.
POV: a 15-second first-person scene with the text doing the storytelling
The POV format in which the camera is the viewer's eyes and one line of on-screen text sets the situation; the prompt writes the text, the single action and the beat where it turns.
Reaction / duet style: respond to a clip in 30 seconds without stealing it
The split-screen or picture-in-picture response format, written so the reaction adds a fact or a demonstration instead of a face. The prompt plans the pause points and what you say at each.
Satisfying process / ASMR: 30 seconds of one process, real sound, no talking
The sound-led format that holds without a hook sentence because the first frame is mid-action and the audio is the process itself. The prompt plans the cuts to the sounds and the single text line.
Screen-recording walkthrough: 45 seconds, one outcome, zoom on every click
The faceless screen walkthrough that stays watchable because every click is zoomed and labelled and the outcome is shown first. The prompt returns the click list, the zoom plan and the voice-over.
Comparison: two ways to do one thing, side by side, same task, same clock
The split-screen comparison that is fair enough to be believed: identical task, identical start, a visible timer, a verdict earned on screen. Prompt returns the script and the fairness checks.
Silent text-on-screen story: 25 seconds, one line per shot, paced for reading
The no-voice format that works with sound off because the text is the narration; the prompt paces each line to its shot at 2 words a second and puts the turn on the second-last card.
'Stitch this' opinion: 5 seconds of their claim, 25 seconds of your answer
The stitch format in which a short borrowed clip sets up a claim and you answer with one argument and one demonstration. The prompt picks the 5 seconds to keep and bans the straw man.
Storytime with a twist: a 60-second told story with the turn at 40 seconds
The talking-to-camera story that holds because the opener promises a turn and the turn arrives late. The prompt plans the beats, the b-roll inserts and the one line of text per beat.
Street interview style: one question, 5 answers, 40 seconds, real or generated
The vox-pop format: one short question, five contrasting answers, the last one lands the point. The prompt writes the question, the answer spread and, if generated, the per-person shot prompts.
'Things I wish I knew': 5 lessons in 50 seconds, each with what it cost me
The advice format that gets saved because every lesson is attached to the mistake that taught it. The prompt orders the lessons by cost and plans one number on screen per lesson.
'3 mistakes' in 40 seconds: the educational format viewers save
Three mistakes the viewer is making right now, each with the fix shown in a shot, ranked so the last one is the one they did not know about. Prompt returns the script, text plan and shot list.
Tutorial in 30 seconds: result first, then the steps at reading speed
The compressed how-to that gets saved: the finished thing in frame 1, then 3-5 numbered steps each shown and captioned. The prompt budgets the words so the steps fit at 165 wpm.
'Watch me use it': a 30-second product demo with no pitch and one real task
The demo that converts because it is a task, not a feature tour: one thing the viewer wants done, done on camera, in real time where possible. The prompt forbids adjectives and plans the text.
Generate 20 hooks for a topic and rank them by a stated rubric
One prompt that produces 20 openers across the 8 archetypes, scores each on four named criteria, and returns the top 5 with the score shown, so the ranking can be argued with instead of trusted.
Hook for a generated video: what must be in frame 1 when a model draws it
The first-frame rules for Veo, Kling and Runway: start mid-action, name the motion already happening, keep text out of the generation, and generate the hook shot as its own short prompt first.
A/B test hooks with campaign links: judge after a week on installs, not views
The creator-growth method: same video, different hook, one store campaign link per variant, installs per campaign read after a week. Includes the iOS ct= link shape and the noise rule.
Bold claim hook: say the thing you can prove in the next 20 seconds
Template and rules for the claim hook: one declarative sentence that the viewer expects you to fail to back up, plus the rule that the proof must be in the video, not in the caption.
Contrarian hook: disagree with the standard advice, then earn it in 20 seconds
Template for the hook that reverses what the audience has been told, with the two rules that separate it from rage bait: quote the standard advice fairly, and name the case where it still holds.
Curiosity-gap hook: withhold one fact, promise it by a stated second
Template for the hook that opens a gap and names when it closes, with the rule that keeps it from being clickbait: the withheld fact is real, single, and delivered where you said it would be.
Number hook: a count the viewer can track, and the one item that justifies it
Template for the hook built on a number (3 mistakes, 5 tools, 7 days), with the rule that the number must be small enough to finish and the last item must be named as the reason to stay.
Pain-point hook: name the exact moment, not the general problem
Template for the hook that opens on the viewer's own bad moment, with the specificity test (a place, a time, an object) that turns 'struggling with X' into a line the right viewer cannot scroll past.
Question hook: the opener the viewer answers in their head before you do
Template and rules for the question hook, with the test that separates a question that holds (the viewer has an answer and wants it checked) from one that is skipped (nobody cares either way).
Result-first hook: show the finished thing in frame 1, say how long it took
Template for the hook that opens on the outcome, the one that works with sound off and for products whose output is the content; the first frame is the result, not you holding it.
'Stop doing X' hook: the correction that must come with the replacement
Template for the interrupt hook that names a habit the viewer has and tells them to stop; the rule that makes it work is that the replacement is in the same sentence or the next one.
Hook rewrite: a boring opener in, 5 stronger versions out, each with the reason
A prompt that diagnoses why an opener is weak (generic, unprovable, needs sound, buried number) and rewrites it five ways, each labelled with the archetype and the fault it fixed.
Hooks for a product with nothing to show: text and voice carry the first 3 s
For an API, a CLI, a service, a habit, anything without a screen worth filming: five hook shapes built on a typed line, a number and a voice, and what goes on screen instead of the product.
Re-hook at 3 s and at the midpoint: two lines that carry a viewer past the drop
Retention is lost twice, after the hook pays off and halfway through; this prompt writes the second promise for 3 s and the third for the midpoint, each as a line the viewer can check before the end.
Visual hook rules: first frame, text on screen, sound off, and how long you have
The checklist for what the viewer sees before any word is heard: first frame, text, the sound-off case, and the attention windows the creator-growth skill states (about 0.1 s, 0.5 s, 1-2 s by age).
Turn a blog post into three short scripts: one claim per video
A prompt that reads a post and returns three 30-second scripts, each built on a different claim from the text, with the hook, the proof, the ask and the source sentence for each.
Forced reasoning before scripting: goal, hook type, demo moments, emotion arc
Four questions an agent must answer in writing before it writes a line of script. Skips the generic promo every model produces when asked for one directly.
The 30-second promo script: hook, problem, demo, CTA with timings
A fill-in script template for a vertical product promo: four blocks with second counts and a 55-90 word narration budget, so the video ends before the viewer scrolls.
Cut any script to 55 words: the 20-second rewrite prompt
A rewrite prompt that compresses a script to 55 words (about 20 seconds at 165 wpm) without losing the number, the name or the ask, and reports what it removed so you can argue.
Screen-recording demo script for an app: what to click, what to say, when
A prompt that writes a demo script where every spoken line is paired with one on-screen action, so the recording is done in one pass and the narration recorded over it afterwards.
Write the cut as a script, not a timeline: the JSON storyboard prompt
An agent prompt that writes the edit of real footage as a small JSON file: which sentences, which cover shots, what the cold open is. A complaint becomes a one-line edit, and the render repeats.
Shot list generator: a table with shot, duration, visual, audio and overlay
An agent prompt that turns any script into a numbered shot table a camera operator or a video model can follow: one visual, one audio line and one text overlay per row, durations that add up.
60-second founder update: the script prompt for a weekly build-in-public clip
A prompt that writes a 60-second talking-head update from your week's notes: the number first, one thing that broke, one thing shipped, one ask. About 150 words, no greeting.
One script, two languages: EN and RU versions that match shot for shot
A prompt that writes a script in English and Russian side by side, same shots and same timings, with the Russian rewritten rather than translated so it fits the same seconds.
YouTube explainer, 8 minutes: the outline prompt with timestamps
A prompt that produces a timed outline for an 8-minute explainer: a cold open, a promise, five sections with a visual each, the mistake, a recap and one ask, before any script is written.
Analyse once: sidecar files beside the sources, shooting time from the container
The storage rule that makes a long film possible a month later: each clip's shots, frame scores, beats and speech live in a file beside it, and shooting time comes from the container metadata.
Cold open, then hook: 8-10 s of best frames, hand-picked from a contact sheet
The opening that holds a viewer on real footage: 8-10 s of the strongest frames with no speech, then one line with a twist. What to look for on the sheet, and why the automatic pick is wrong.
The critic pass: repeats and sagging scenes first, mid-word joins second
A two-pass review prompt for a subagent or a second session: pass one on structure (repeats, sagging scenes), pass two on joins (mid-word cuts, audio seams). One read finds one kind of problem.
A contact sheet of every cutaway with its line under it
The one-page check that catches off-topic cutaways before anyone watches: every cutaway frame in a grid, its sentence under it. Measured once: 12 of 54 automatic cutaways were off topic.
The cutaway rule: what is said, same place and time, never over moving lips
Three constraints for every cutaway in a talking film, and the agent prompt that applies them to a transcript. Kills what viewers read as broken: a picture from elsewhere, a face out of sync.
The day as one text file: an agent reads the speech and proposes the story
The step before any timeline: the day's speech in shooting order, numbered by clip and sentence, read by an agent like a scriptwriter. It proposes story, scenes, hook and cutaways; you argue in text.
Find the cuts in a video with ffmpeg's scene filter
One ffmpeg command that lists the timestamps where the picture changes, how to set the threshold, and why phone footage often has no cuts to find.
Length follows the material: cut stammers and repeats, not thoughts
The decision rule for how long a film should be. A target length makes you stretch a thin day or chop a full one; the material decides, and the checklist says what may be removed.
Promise and payoff: find the pairs that thread a multi-day film
An agent prompt that reads several days of transcripts and pairs every promise ('tomorrow we'll buy the boat') with the moment it is kept or broken, so the film has a thread, not a sequence of days.
Read the film as text before rendering: the last check that costs nothing
The cut, printed as its lines in order, read by someone who never saw the raw footage. Catches the question whose answer was cut, the name said before the person, the 'this' that points at nothing.
Scenes, not best phrases: keep the lead-in, the talk and the reaction
Why a reel of good lines confuses everyone who was not there, and the checklist for keeping a scene whole: the lead-in, the talk, the reaction, with filler trimmed only at the ends.
Take something from every clip, and list the ones you cut with a reason
A rule for cutting someone else's footage, or your own from a year ago: each clip was filmed because it mattered, so each gets a chance, and the rejects get a written reason the person can argue with.
Voice as one continuous track, pictures on top, and the one-frame length check
The assembly rule for talking films: cut the audio from the same pieces as the picture, snapped to frames, then verify film length equals voice length. Drift over one frame is a bug, not a quirk.
Burned subtitles for 9:16: position, lines, characters, colour, as parameters
Settings for subtitles burned into vertical video for muted feeds: above the platform UI, two lines, about 32 characters each, white with a dark outline. Starting values, checked on a phone.
Accessible on-screen text: contrast, the bottom 20 percent, no flashing
A checklist for text that everyone can read, including people watching muted, on a small screen, in sunlight, or with photosensitivity. Eight checks, each with the parameter to test against.
Three caption variants plus hashtags, with per-platform length limits
An agent prompt that writes three captions (short, story, question) and a hashtag set per platform, counting characters and hashtags against limits you paste in from today's upload form.
Burn subtitles with ffmpeg: the command with size, outline and margin
One ffmpeg command that burns an SRT into a 9:16 video with white text, a dark outline and the block held above the platform UI, plus the check that your ffmpeg build can run it at all.
Hook text on screen: a plan for the first three seconds, word by word
A prompt that plans what text appears in the first 3 seconds of a vertical video: the words, the moment each appears, the hold, what the picture does under it. Written so the hook works muted.
Reading speed for on-screen text and speech: the parameters to check against
How long a text overlay must stay up and how many words a voice can carry per second, given as parameters you measure on yourself, with the formula and the two failure modes: unreadable, and slow.
SRT cleanup: fix names from a glossary, merge short cues, cap line length
An agent prompt that takes a raw speech-to-text SRT and returns one you can publish: names fixed from your glossary, sub-second cues merged, lines wrapped at a character cap, timings untouched.
Transcript to YouTube chapters: timestamps in the format the description needs
A prompt that turns a timestamped transcript into a chapter list for the YouTube description: first chapter at 0:00, at least three, each 10 seconds or more, titles that say what the viewer gets.
From a sentence to five search queries and one generation prompt
A prompt that turns each line needing a cutaway into five queries for your own footage or a stock site, ordered from literal to loose, plus one generation prompt as the fallback when nothing is found.
B-roll shot list from a transcript: the table to film or generate from
An agent prompt that reads a transcript and returns a shot table (timestamp, line, shot, source: have / film / generate) for a camera or a video model, grouped so each location is filmed once.
Cutaway matching: sentence to a literal visual, same place and time
A prompt that reads each sentence of a transcript and names the one picture that shows it, with the place and time it must come from, so the shot list is a list of things, not moods.
Generated b-roll or real: the decision rule
When a generated cutaway is fine (faceless, no logos, no claim about a real place or time) and when only footage from the same day will do. Five questions, in order, before you open a video model.
A mentioned object must be seen, ideally just before the line about it
The rule that turns 'X is mentioned, never shown' from a viewer complaint into a check you run on the transcript: every named thing gets a shot, placed just before or on the line.
Screen-recording b-roll for app demos: what to record, size, cursor
Rules for screen captures used as cutaways in a product video: real-looking data, output-size or 2x resolution, a slow cursor, one action per clip, and what makes a recording unusable.
Recording to MIDI: pick the model by material, then check it
Three local transcription models fail in different directions. The skill picks one by what the material is, runs it, and checks the result before you trust it.
Beat-sync cutting: cut on the beat, not on every beat
A rule and an agent prompt for cutting to music from a list of beat timestamps: cuts land on beats, phrases get room to breathe, and the best shot is held through the hook.
Prompt for a music generator: a 30-second instrumental for a short
A brief in the shape music generation tools read: structure by seconds, BPM, three instruments, no vocals, a drop that lands on your hook cut, and an ending that stops instead of fading.
Live music in frame: one continuous audio span, picture cuts on top
When a band or a busker is in the shot, keep one unbroken stretch of that clip's audio under the whole scene and cut only the picture; cut the audio per shot and the beat restarts.
Level every chapter to -16 LUFS with ffmpeg loudnorm
One ffmpeg line that brings a file to -16 LUFS integrated and -1.5 dBTP true peak without touching the video, the two-pass version for exact results, and the timestamp trap on upload.
Compare two microphones by numbers, not impressions: mic-compare
Record the same phrase on two mics at matched gain and let the agent report response, noise floor, per-band SNR, plosives and sensitivity, and what EQ can still fix.
Music brief: describe the track you need before you search for it
A fill-in brief that names mood as harmony, a BPM range, genre, length and where the drop must land relative to your hook, so a library search or a composer returns something that fits the cut.
Score with a hole, not ducking, where something plays live
Cut the score to silence under live music, a laugh or a knock instead of ducking it; ducking is audible, a hole is not. The rule, the measured levels and the ffmpeg line.
Render plugin chains from a text file, no DAW: sdsp-chain
One TOML file describes tracks, effects, sidechains, automation and playable instruments; a Rust command renders it to WAV and prints LUFS and true peak per stage.
Mashups and ciphers with a beat grid: sdsp-mash
Pick pairs by BPM and key, split stems, stretch only the beat, and render one beat with verses from other songs, or a beat switch per verse, from a TOML file.
Mix and master by measurement: sdsp-mix
An agent mixes the way an engineer would: render, measure against a commercial reference, change one thing. Mix map of who masks whom, kick and bass ducking, mastering to a loudness target.
Strip other people's music before upload with Demucs and ffmpeg
Car radio, shop speakers and festival stages trigger Content ID. Split the soundtrack into stems, find the music on drums plus bass, keep the voice, replace the music, rebuild audio only.
Nine agent skills for mixing, mastering and making music (superduper-dsp)
A Claude Code plugin that teaches an agent to mix by measurement, build songs in REAPER, render plugin chains headless, turn audio into MIDI and tear down a reference track.
Build a track in REAPER with an agent: superduper-song
The agent writes the song into REAPER through its MCP bridge on the SuperDuper instruments, with parameter maps by name, preset recall and expression via CC.
Tear down a reference track by numbers: tempo, key, groove, balance
The agent reads a spectrogram instead of listening, splits stems with Demucs, fits tempo and phase to the kick, and rebuilds the vibe until the numbers match the reference.
A vocal chain slot by slot: plosives, gate, compress, air, de-ess
The vocal-mix skill lays out a nine-slot channel strip in REAPER with the reasoning for the order. The numbers are tuned to one voice and one mic; the order carries over.
Voice cleanup checklist: room tone, de-ess, noise gate, in that order
Six passes that turn a phone or lav recording into a usable voice track, each with the ffmpeg filter and a starting parameter to adjust by ear, in the order that keeps the gate off word endings.
Measurement: one store campaign link per platform and account
The iOS campaign link format with the provider token and a 40-character ct tag, the Android Play referrer, where installs show up, and the rule to judge a format after a week on installs, not views.
The creator budget gate: can you afford a paid creator yet?
A fill-in prompt that divides a creator's monthly retainer by your net revenue per sale to get payback sales per month, then says whether you are in Phase 1 (founder-led, $0) or allowed to pay.
The download ask lives in the comments, not in every video
A hard sell in every video gives viewers a reason to scroll. Name the app in replies to 'what app is this?', keep the link in the bio, and write the replies before you need them.
Phase 1, founder-led: the four formats and what each one buys
Educational converts, POV and humor get reach, story builds trust, comparison works when the app replaces something expensive. A prompt that plans a week of founder videos across all four.
In-app growth levers: brand credit, 9:16 output, one-tap preset
Four product changes that turn every export into UGC and make a creator's demo work in one tap. An audit prompt for the codebase that marks each lever done or next, plus the disclosure rule.
2-3 niche accounts, one audience archetype each, one app per account
Why several small accounts beat one big one for an app: each speaks to one kind of viewer. A prompt that defines the archetypes and the account plan, and the clone trap that gets accounts suppressed.
Phase 2 trigger: when to pay micro-creators, and how much
Pay nobody until one video passes a view threshold AND its campaign link shows installs. Then 3-5 micro-creators in the same niche, paid per video or per million views, no retainer yet.
Cadence: 1-2 posts a day, one vertical cut to TikTok, Reels and Shorts
The minimum posting rhythm for Phase 1 and the production habit that makes it possible: one 9:16 master per idea, three uploads, no per-platform re-edit. With the ffmpeg line for burned captions.
Weekly growth review: an agent reads the posts and the installs per campaign
Paste the week's posts and the installs per campaign link; the agent reports numbers first, then what worked by name, then 3-4 actions, judging by installs and by multiples, never by a 20 % swing.
YouTube upload checklist: gate the file, publish private, watch on a phone
Cheap checks first, upload last: title under 100 characters, the real shooting date and chapters in the description, captions as tracks, Shorts linked both ways, a private publish watched on a phone.
Beat times and downbeats from a song in Node, with life2film-engine
A short Node script that decodes a song with ffmpeg and prints BPM, beat times and downbeats on your machine, ready to place cuts on. No upload, no API key.
Scan a screen recording for leaked secrets before you post it (screenshield)
Before a demo, lecture or stream goes public, scan the video for API keys, personal data and QR phishing, and blur what it finds. On-device OCR, no LLM needed for the scan.
Turn an article or PRD into a social media pack (/solo-content-gen)
Turn an article or a PRD into a social pack: LinkedIn posts, short captions for Reels and TikTok, and a thread, with claims marked as sourced or not.
Plan short-form growth for an app: founder-led first, creators after a signal (/solo-creator-growth)
Plan growth for a consumer app through short vertical video: founder-led content first, paid creators only after a measured signal.
Strip AI patterns from a script or caption so it reads like a person (/solo-humanize)
Strip stock AI phrasing from a script or caption: em dashes, triplets, filler openers, the hedging sandwich. Numbers, names and caveats stay.
Write a landing page with copy, SEO, JSON-LD and a cookieless counter (/solo-landing-gen)
Turn a PRD into a landing page: A/B headlines, section copy, a type and spacing system, JSON-LD, llms.txt and a cookieless counter.
Cut a vertical reel to a song with real beats (/solo-music-reels-cut)
Footage plus a music track becomes a 1080x1920 reel: beats and phrases detected for real, cut density that follows verse and chorus, 14 ffmpeg effects, a flash on every downbeat.
Separate vocals, music and drums from a recording (/solo-stem-split)
Separate a song or a film's soundtrack into stems (vocals, drums, bass, other) with Demucs, locally, on the Apple GPU when there is one.
Listen to a live room and keep one running summary (/solo-supervox)
With supervox transcribing locally, the agent keeps one evolving summary on screen during a talk or a call, coaches a pitch's Q&A, or analyses a saved recording afterwards.
Transcribe audio and video locally with word timings (/solo-transcribe)
Transcribe audio or video to text locally: on the Apple GPU with mlx-whisper large-v3, or faster-whisper on CPU/CUDA: with the language decided per stretch of speech, .srt subtitles, word timings…
Edit real footage into a film with the agent as the editor (/solo-video-edit)
Edit real footage into a film with the agent as the editor: read the day's speech as text, write the cut as a script, lay cutaways that match the words, score it, strip other people's music, and…
Write a 30-45 s promo video script and storyboard from a product (/solo-video-promo)
Plan a 30-45 second vertical promo from a README or PRD: goal, hook type, three demo moments, a timed script and a storyboard with an assets checklist.
Upload, title, chapter and track videos on a YouTube channel (/solo-youtube-channel)
Run a solo founder's YouTube channel from the terminal: upload through the YouTube Data API, subtitles as tracks, thumbnails, publish, bulk hide, link Shorts to the long video, and read the…
What to do next
- Turn an idea into a ready Veo prompt: Veo 3 prompt generator or the video prompt builder. Both run in your browser.
- Cut footage you already shot: Life2Film scores the frames and builds the montage on your iPhone or Mac, nothing uploaded.
- Let an agent do the editing: install the skills with
npx skills add fortunto2/solo-factory --skill solo-video-edit, see Skills & tools. - Follow new entries: YouTube, Instagram, GitHub. There is no newsletter.
Longer reads and free tools from our other sites
- Edit video with an AI assistant: Claude Code, ChatGPT and Codex · life2film.com
- How to cut video to the beat, by hand and automatically · life2film.com
- Edit video with an AI assistant: a day of footage to a film · rustman.org
- How the Chicago films were actually cut: an agent, a Rust core and ffmpeg · rustman.org
- Speech is the timeline: transcription for video editing · rustman.org
- Sound for travel films: your own score and no radio · rustman.org
- A cut detector found zero cuts in a day at the beach · rustman.org
- My beat tracker missed nine downbeats in ten · rustman.org
Free tools: BPM detector (runs in the browser) · Scene detector (runs in the browser) · life2film-engine on npm
How this library works
- Every entry is one prompt, recipe or skill, written to be pasted as is. Placeholders are in {braces}.
- "Works with" is a promise about where the prompt was written for, not a claim that we tested every combination. Verified appears only on entries we ran and looked at.
- Claude Code skills and tools come from solo-factory (MIT) and install with one line.
- Generation prompts route to a video model; editing recipes work on real footage, by hand or with Life2Film on your device.
- For agents: library.json, llms.txt, and a
.mdtwin for every entry. 0 of 135 are marked verified. - Wrong, outdated, or you have a better one: [email protected].