Ten ideas you could film this week
Describe the channel in a few sentences and pick what you want the video to do - reach, followers, sales or authority. You get ten specific ideas with a working title, the first sentence to say out loud and why it should work for your audience, plus three series concepts that turn one good video into ten.
The AI Video Idea Finder suggests video ideas written for a channel goal and a target audience, each with a filmable angle, instead of a generic list of topics. Ask again with a different goal to get a new set. The ideas are generated by an AI model and need a check against what you can film.
The idea drought is a filter problem
Almost nobody runs out of topics. What runs out is the patience to throw away the nine ideas that are interesting in a notebook and impossible to film in a living room on a Thursday evening. A list of twenty vague directions feels like progress and produces nothing, because the next step is never obvious: which one, in what format, with what first sentence.
A usable idea has four parts, and a missing part is what makes an idea feel heavy. It has a subject narrow enough to be answered in one video. It has a reason that someone is searching or scrolling for exactly this. It has a format that matches the reason - a tutorial for a how-to, a list for a comparison, a story for a lesson that only makes sense as a story. And it has a first sentence, because the first sentence is the part that decides whether any of the rest is watched.
The tool is built around those four parts. It is not a random idea machine: it takes the niche you describe, the platform's format and the goal you choose, and it writes the four parts together so the idea arrives ready to film. Ten of them at once is deliberate - the point is to have enough that you can reject seven without feeling the loss.
Effort buys reach, but only in some formats
The two axes below are the only ones that matter when you are choosing what to film next: how much work the video needs and how far it travels after publication. Every format sits somewhere specific, and a channel that only lives in one corner eventually stops growing.
A channel that only lives in the top-left corner grows until the easy ideas run out. The two right-hand formats are what make the next hundred videos possible.
The top-left quadrant is where most creators start and where the first ten thousand views come from: answers to questions people are already typing. It is also the quadrant that runs out, because a niche only has so many questions. The right-hand column is where a channel gets a personality, and it is expensive: a case study needs a real result, an origin story can only be told once. The two bottom-left formats are the cheap experiments - useful for testing a style, dangerous as a diet.
What each format costs and how long it lives
| Format | What it needs | How long it keeps working | Best goal |
|---|---|---|---|
| Tutorial | a process you can show step by step | years, and it is found by search | authority, followers |
| List | five to nine items you can defend | months, until the list is out of date | reach |
| Story | a real sequence of events with a turn | one strong week, then it is a reference | loyalty |
| Hot take | an opinion you can argue in four minutes | days, and it invites comments | reach |
| Case study | numbers, a before and an after | years, and it converts | sales, authority |
| Reaction | something that already happened | a week at most | reach, community |
The column that matters for planning is the third one. A channel that publishes only hot takes has to publish constantly, because every video dies on schedule. A channel with a body of tutorials keeps getting views from videos published two years ago - which is the difference between working for the algorithm and having an archive that works for you.
From the list to a filmed video
Cut to three
Take the three ideas you could film without buying anything. Everything else is a maybe, and maybes do not get filmed.
Say the hook out loud
If the first sentence is awkward spoken, the video will be too. Rewrite it before you write anything else.
Fix the format
Decide the shape before the script: five steps, one story, three examples. The shape is what stops the middle from sagging.
Put a date on it
An idea without a publish date is a note. Put the three in the calendar and film the easiest one first.
Where ideas actually come from
The best source is the least glamorous: the questions under your own videos and the ones people send in private. A question asked three times is a video, because the fourth person is already searching for it. The tool cannot read your comments for you, but it can turn "how do I keep the bottle cold in summer?" into a titled video with a hook.
Paste the questions into the niche box as they were written. The phrasing of a real question is worth more than a polished brief, because it is the language the audience already uses.
Open the five channels closest to yours and sort their videos by popularity. What the top ten have in common is the demand; what is missing between them is the opportunity. A topic covered badly by three channels and never from your angle is a better bet than a new topic nobody searches for.
When you describe your niche to the tool, mention what is already covered. "Everyone does beginner guides, nobody covers what to do in month two" is a sentence that changes every idea it writes.
The three series concepts at the end of the result are the part most creators ignore and the part that compounds. A series gives you a reason to publish next week, teaches the audience to come back, and makes the playlist a page that ranks on its own.
A series needs a promise the viewer can repeat: "one tool every Tuesday", "a teardown per month". If you cannot say the promise in one sentence, it is not a series yet.
Choosing what to film, without pretending it is random
Demand first, personality second
Two kinds of video bring an audience, and they do different jobs. Search-driven videos answer a question that already exists: how to do something, which one to buy, why something broke. They are found months later, they bring subscribers who stay, and they are boring to make. Feed-driven videos earn attention in the first second and disappear in a week: a take, a reaction, a story. They build the relationship that makes the search videos get watched when they are recommended.
A channel that publishes only one kind eventually stalls. All search and no personality gives you views without a following; all personality and no search gives you a following that cannot find you again. The honest ratio for a channel under a hundred thousand subscribers is roughly two search videos to one feed video, adjusted by however much you enjoy talking to camera.
The test that kills bad ideas early
For each idea, answer three questions in one sentence each. What does the viewer know after watching that they did not know before? Why would they click on this thumbnail rather than the one below it? Can you film it with what is in the room right now? An idea that fails the third question is not necessarily bad, but it is not this week's video, and treating it as one is how channels go quiet.
The second question is the one creators skip, and it is the one the platform judges. A title is a promise and a thumbnail is the evidence; if the two are not in tension - the promise is surprising, the evidence is specific - the video is only as good as the algorithm's mood.
How to use ten ideas without filming ten videos
Treat the list as raw material. Group the ideas that share a shoot: three ideas about the same tool can be filmed in one session, and the footage can serve a long video and two shorts. Look for the idea that is really a series, and for the one that is really a section of a bigger video. The tool gives you the seed; the schedule decides what actually gets made.
Keep the rejected ideas. A month later, the one that felt too obvious will look like the only thing worth making, and the one that felt brilliant will look like work. The list is not a to-do list - it is a record of your own taste, which improves only if you can compare what you liked with what worked.
Platform differences that change the idea
The same idea is a different video on each platform. On YouTube the first thirty seconds carry the promise and the search title carries the discovery; on Shorts and TikTok the hook is the video and the loop is the ending; on Reels the first frame has to work without sound. Choosing the platform before the idea is why so many good topics die in the edit - the format was decided by habit, not by the idea.
A word on the daily limit
Each run costs money on the API, so the tool gives five free runs per day per visitor. That is enough for a monthly planning session: one run for the channel, one for the audience you are trying to reach, and three for the specific series you are unsure about.
Before you write the script
Are the ideas original?expand_more
Why does it ask for a goal?expand_more
Can I use one idea on several platforms?expand_more
What if the ideas are too broad?expand_more
Does it write the script too?expand_more
Ten ideas, one calendar
Describe the channel once, pick the platform and the goal, and leave with a list you can film.
fact_checkQuick facts
- An idea is a promise plus a format: the same topic asked as a tutorial, a story and a comparison becomes three different videos.
- Ideas are cheap and filming time is not, so the useful test is whether you can shoot the idea with the equipment and the access you have today.
- A language model reads text in tokens, and one token is roughly four characters of English, so a 300-word answer is about 400 tokens.menu_bookOpenAI Cookbook
- The output depends on the prompt: the same request with a different tone or length instruction produces a different text, not a different answer quality.menu_bookOpenAI Cookbook
- YouTube defines impressions as the times a thumbnail was shown, and the click-through rate as the share of those that became views.menu_bookYouTube Help
helpQuestions people ask
How do I come up with video ideas for my channel?expand_more
How many video ideas should I make in one batch?expand_more
Can an AI pick a video idea that will go viral?expand_more
hubWhat this page is about
- Large language modelopen_in_new
A model that predicts the next token of a text; the site sends the prompt to a hosted model and returns its answer.
- Click-through rateopen_in_new
Clicks divided by impressions, the percentage of people who saw a thumbnail and chose to click.
library_booksOfficial sources cited on this page
- How to count tokens with tiktokenOpenAI Cookbook
- Thumbnail impressions and click-through rateYouTube Help
- Custom video thumbnailsYouTube Help
Keep going with the next tool
The tools in this group that do the next piece of the same work, all switched on right now.
Switched-off tools are not listed here: only what works right now.