> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pictory.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Create Videos with the Skill
> How to ask for videos, what the skill does with your request, tips for great results, and troubleshooting
With the skill [installed and configured](/claude-skills/text-to-video-setup), creating
a video is a conversation. This page shows what to say and what to expect.
## The basic flow
Start Claude Code and describe the video you want:
> Create a 45-second video about the benefits of morning workouts. Energetic voice,
> upbeat music, vertical format for Instagram Reels.
Here's what happens next:
1. **A few quick questions.** If you left out something important (format, voice-over,
music, whether AI-generated visuals are okay), the skill asks once, all together.
Answer, or just say *"you decide."*
2. **A scene plan.** For longer videos you'll see a short outline first: each scene's
narration, visual idea, and layout. Say *"looks good"* or ask for changes. This is
the cheapest moment to redirect.
3. **Rendering.** The skill submits the video to Pictory and posts a progress update
every 30 seconds. Most videos finish in a few minutes; AI-generated visuals take
longer.
4. **Your video.** You get the direct MP4 link, a shareable player link, and (by
default) an editable copy saved to **My Projects** at
[app.pictory.ai](https://app.pictory.ai) for hand-polishing in the Pictory editor.
## Example prompts
**Marketing & social**
> A 30-second launch teaser for our smart water bottle. Premium, futuristic look, dark
> background, AI-generated product shots. End with "Pre-order now."
> Three-scene vertical short: why remote teams need async updates. Casual tone, bold
> captions, trending-style pacing.
**Explainers & tutorials**
> A 90-second explainer on how two-factor authentication works, for non-technical
> employees. Friendly voice, simple language, calm music.
> A tutorial showing the three steps of our onboarding flow as a connected diagram that
> lights up step by step. 16:9 for our help center.
**From your own material**
> Turn this blog post into a one-minute video with subtitles: [https://example.com/post](https://example.com/post)
> Here's my script. Keep every word exactly as written, you design the visuals:
> \[paste script]
## Tips for great results
* **Say who it's for and where it will play.** "For LinkedIn" or "for our help center"
changes the format, pacing, and tone the skill picks.
* **Give a mood, get a design.** Three adjectives like "warm, playful, personal" or
"urgent, technical, confident" steer the colors, fonts, music, and visuals.
* **Mention brand colors or a logo URL** if you want them used.
* **Length matters.** Ask for 15-20s for a teaser, 45-60s for a standard video, 90s+
for a deep dive. Narration is paced to fit; a tighter video is usually a better one.
* **AI visuals are opt-in.** Stock footage is free with your plan; AI-generated images
are cheap; AI-generated *video* uses meaningful AI credits. The skill always asks
before spending credits and tells you roughly what it used afterward.
* **Iterate freely.** After the first render: *"Make scene 2's headline shorter"*,
*"different footage for the intro"*, *"swap to a female voice"*. The skill edits the
same video and re-renders.
## What you get back
| Output | Description |
| ---------------- | ------------------------------------------------------------------------ |
| Video URL | Direct MP4 download link |
| Share URL | A hosted player page you can send to anyone |
| Thumbnail | A preview image |
| Editable project | The video in **My Projects** at app.pictory.ai, ready for manual editing |
| Captions files | SRT/VTT subtitle files for re-use |
## Troubleshooting
Set `PICTORY_API_KEY` as shown in [Install & Configure](/claude-skills/text-to-video-setup),
then open a **new** terminal and restart Claude Code. Check with:
`echo ${PICTORY_API_KEY:+set}`. It should print `set`.
The key is wrong, expired, or your API subscription isn't active. Copy the key again
from the [API Subscription](https://app.pictory.ai/api-access) page. Note the key is
sent as-is; if you're adapting the calls yourself, don't add a "Bearer" prefix.
A video only gets a voice-over when one is requested. Say "add a voice-over" (and
optionally which voice) and re-render. The skill includes one by default unless you
asked for a silent video.
Say "lower the music". The skill sets music volume explicitly (around 10-15% under
narration) and will re-render with your preference.
Renders with AI-generated visuals can take a while. The skill polls for up to 30
minutes and reports progress every 30 seconds. If it times out, it gives you the job
ID. Ask it to check that job again later, or find the video in My Projects once
it finishes.
AI-generated visuals need AI credits on your Pictory plan. Ask the skill to use
stock footage instead, or top up credits at app.pictory.ai.
Skills load when a session starts. Make sure the folder is at
`~/.claude/skills/pictory-text-to-video/` (or `.claude/skills/` inside your
project), then restart Claude Code and type `/pictory-text-to-video`.
## Related
The underlying API, if you'd rather integrate directly
Use Claude Code with the full Pictory API beyond this skill