The InVideo AI video generator combines project-level AI assistance, media generation and a timeline editor. The useful starting question is whether you need a whole sequence planned, a new shot generated, or existing footage assembled. Those are different jobs, with different opportunities to spend credits unnecessarily.
For a character you also want to develop and chat with, Cherrypop’s character creator serves a more specific task. For a general video edit with a soundtrack and multiple cuts, the timeline workflow below is the relevant starting point.
Choose your InVideo AI video generator workflow
| What you have | Route to consider | Decision to make first |
|---|---|---|
| An idea but no sequence | Agent-led project | Which scenes need to exist, and what should the viewer understand? |
| One image that needs movement | Image-to-video through Agents & Models | What visibly changes during this single shot? |
| Usable footage and audio | Timeline Editor | Which cuts and sound changes finish the sequence without regenerating footage? |
InVideo’s filmmaking introduction describes Agent mode as the prompt-page starting point for a script, concept or scene. Its free-access guide separately documents image input for video generation through Agents & Models. Treat those as two entry routes, rather than assuming every project begins with the same model settings.
A useful project brief names the audience, the required events and the assets you already own. “A short video about a ceramic lamp” leaves many decisions open. “Show the lamp on a desk, then the switch being pressed, then the illuminated shade” establishes three visible events you can evaluate individually.
Allocate the scene before generating it
Suppose you want a short showcase for a fictional desk lamp. Here is a proposed three-shot plan. It avoids asking one generation to establish the room, demonstrate a mechanism and deliver an accurate final product view all at once.
| Shot | Required information | Best available input | Reject when |
|---|---|---|---|
| Opening view | Where the lamp sits and how it looks | Your approved still or existing footage | The base or shade differs from the intended design |
| Action | A clear change from off to on | A short controlled action request or real footage | The switch changes position or the action is unreadable |
| Closing view | The accepted lit appearance | A usable final frame or separate still | The object no longer matches the opening |
The editorial benefit is replacement scope. If the action fails, replace that shot rather than rebuilding the whole sequence. If the lamp already has a good closing still, a hold on that image may communicate the result without paying for unnecessary new motion.
For a real product, use an accurate source and inspect every identifying detail. A plausible fictional lamp is not evidence that the tool can faithfully demonstrate a real switch or certified lighting effect. This example illustrates planning, not measured physical accuracy.
Move from a rough project to a controlled edit
The Editor introduction documents a browser timeline with AI assistance and manual controls. To start with your own assets, select Timeline Editor below the prompt box, create a blank timeline and import media. You can ask the agent to assemble a base cut, then review the actual clips and cuts.
InVideo’s timeline guide separates visual tracks from audio tracks and provides a Monitor for previewing the playhead position. Use that separation to diagnose a rough result: a mistimed cut, an overpowering music track and an inaccurate generated shot require different repairs.
- Arrange the lamp’s three events in a readable order. Check whether each adds information before adding transitions.
- Watch the sequence without music. If the action is unclear, sound will not make the visual demonstration accurate.
- Review sound separately: narration must remain understandable, and a decorative effect should not imply a feature the lamp does not have.
- Play across each join. Change timing manually where timing is the problem; reserve another generation for missing or unusable imagery.
Separate credit spend from finishing work
The plans and credits guide says manual editing and exporting do not consume credits. AI assistance and generated media do. Video generation cost varies with model, duration and settings; it is not a single fixed price per finished video. Paid monthly allowances reset without rolling over, while separately purchased credits have different expiry rules.
That distinction changes the next move after a disappointing result. A cut that lasts too long can be shortened manually. A subject that changes shape needs another visual solution. Paying an agent to rearrange footage and paying a model to create footage belong in separate budget lines.
| Budget line | Record | What it tells you |
|---|---|---|
| Generated media | Displayed charge for completed outputs, including those you choose not to keep | The cost of obtaining usable source material |
| AI-assisted editing | Agent charges associated with assembly and revisions | The cost of delegated editing decisions |
| Manual finishing | Your time spent adjusting cuts, sound and titles | Work remaining after generation |
| Accepted deliverable | Whether the exported file meets the brief | The useful result, rather than the number of previews |
The free-access guide describes limited credits and a weekly reset. It does not justify assuming that every model, output setting or complete project fits inside your account’s current allowance. Read the displayed charge before each generation and choose a small first task.
Export the accepted version, then inspect the file
InVideo’s download instructions place exported videos in the You section: select the export, then Download.
Open the resulting file outside the editor. Check the beginning and ending, the actual crop, audible narration and any visible marks. A good timeline preview does not establish that the saved file meets a client’s dimensions, licensing or delivery requirements. Resolve those requirements before making a large batch.
When a character-centered workflow is a better fit
Cherrypop is useful when the continuing character matters as much as the clip. Its creator lets you define appearance and personality, then continue into character interaction and media creation. That gives an original fictional character a purpose beyond a single exported sequence.
Start with the character identity and scenario, then use Cherrypop’s video generation entry for the media task. The current interface includes image-to-video and continuation modes. We are not presenting it as an InVideo-style multitrack editor or promising matching output quality, duration or export controls.
Cherrypop is free to start with limits; relevant media can require Premium or Cherries. Check the displayed cost before submitting. If the project is an original companion character, create the character first and make the next clip serve that character.