With Narration Flow
What is the narration flow?
The With Narration flow is ReelBot’s most powerful and commonly used creation path.
It combines:
- structured messaging
- AI voiceover
- word-level caption synchronization
- visual storytelling
This flow is designed for clarity, pacing, and retention — not raw automation.
Overview of the steps
The narration flow follows five deliberate steps:
- Topic
- Script
- Voiceover
- Assets
- Review & Generate
Each step:
- has a single responsibility
- builds on the previous step
- can be regenerated independently when needed
Step 1: Topic
The Topic step defines what the video is about.
You can:
- write your own topic manually, or
- generate AI topic suggestions
AI topic suggestions
When generating topics:
- you provide keywords or context
- ReelBot returns multiple topic options
- you choose one to continue with
Only one topic is active at a time.
Once selected:
- the topic is locked
- the next step becomes available
Step 2: Script
The Script step defines what will be said.
You can:
- write your own script, or
- generate an AI script based on the topic
Script limits and duration
Script length is automatically constrained by:
- the selected video duration
- natural spoken pacing
This ensures:
- scripts fit the final video length
- delivery feels natural, not rushed
Regeneration
If the script doesn’t feel right:
- you can regenerate the script
- tone and duration settings are respected
- assets and voice remain untouched
Step 3: Voiceover
The Voiceover step defines how the script is delivered.
In this step, you:
- select a voice
- review voice type and accent
- generate the voiceover
Once generated:
- a preview player becomes available
- spoken pacing is locked
- captions timing is derived from speech marks
Why voice comes before visuals
In narrated videos:
- voice controls timing
- captions follow the voice
- visuals adapt to delivery
This order ensures accurate synchronization.
Step 4: Assets
The Assets step defines what is shown visually.
You must select:
- at least one B-roll video
- optionally, background music
B-roll sources
You can choose from:
- your uploaded videos
- public library videos
- AI-generated animated images
Selected assets:
- can be reordered
- are previewable
- are adapted to the chosen orientation
Music (optional)
If selected:
- music is mixed under the voice
- volume is balanced automatically
- only one track can be used per video
Step 5: Review & Generate
The Review step summarizes all decisions.
You’ll see:
- format and duration
- tone and caption settings
- selected voice
- asset count
- music selection
From here, you can:
- generate the video
- go back and adjust specific steps
Video generation
When you click Generate Video:
- ReelBot composes the video asynchronously
- progress is tracked automatically
- the system polls until completion or failure
On completion:
- the video is saved as a Project
- preview and download options appear
- publishing options become available
Regeneration rules
ReelBot enforces safe regeneration.
- Regenerating script does not affect assets
- Regenerating voice does not affect script
- Changing duration or tone may require clearing earlier steps
You are always warned before steps are cleared.
Draft behavior
The narration flow supports automatic drafts.
- progress is saved continuously
- you can leave and resume anytime
- drafts are removed once a project is created
Drafts ensure creation is never interrupted.
Common mistakes to avoid
- Changing duration after generating voice
- Regenerating everything instead of one step
- Over-optimizing visuals before the message is right
The flow exists to prevent these mistakes.
The CreatorOps perspective
The narration flow is not a form — it’s a pipeline.
By locking decisions in the right order, ReelBot ensures:
- consistent pacing
- accurate captions
- predictable results
- safe iteration
This is CreatorOps applied to storytelling.
What to explore next
👉 Learn how visual-only videos work
→ Cinematic / Music Only Flow
👉 Or explore Video Settings & Templates
→ Video Settings Sidebar
Narration is where clarity is built — everything else supports it.