Skip to content
Back to skills

Text To Video

ASecurity

Generate videos from text descriptions using each::sense AI. Create short video clips, animated scenes, cinematic shots, and motion content from text prompts. Supports photorealistic, animated, and stylized video generation with camera movements and transitions. Use for: video content creation, social media clips, concept visualization, storyboard animation, motion design, creative exploration. Triggers: text to video, generate video, create video, ai video, video generation, txt2vid, prompt ...

  • 29 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added June 2, 2026
content-marketingpythongobashapidocumentation

Works with

  • cli
  • api

Security analysis

A100/100

Scanned June 2, 2026

npx -y skills add awesome-genmedia/skills --skill text-to-video --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Text To Video?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Text To Video
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/awesome-genmedia-text-to-video/badge)](https://www.skillsdirectory.com/skills/awesome-genmedia-text-to-video)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: text-to-video
description: "Generate videos from text descriptions using each::sense AI. Create short video clips, animated scenes, cinematic shots, and motion content from text prompts. Supports photorealistic, animated, and stylized video generation with camera movements and transitions. Use for: video content creation, social media clips, concept visualization, storyboard animation, motion design, creative exploration. Triggers: text to video, generate video, create video, ai video, video generation, txt2vid, prompt to video, ai animation, video from text, ai clip, motion generation"
allowed-tools: Bash(curl *), WebFetch
---

# Text to Video

Generate videos from text descriptions using [each::sense](https://docs.eachlabs.ai/sense/overview) — the intelligent AI agent that automatically selects the best model for your request.

## Quick Start

> Requires an each::labs API key. Get one at [eachlabs.ai](https://eachlabs.ai).

### Using curl

```bash
curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a video: A golden retriever running through a sunlit meadow of wildflowers, slow motion, cinematic, shallow depth of field, warm afternoon light, camera tracking alongside the dog"
      }
    ],
    "stream": false
  }'
```

### Using Python (OpenAI SDK)

```python
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_EACHLABS_API_KEY",
    base_url="https://eachsense-agent.core.eachlabs.run/v1"
)

response = client.chat.completions.create(
    model="eachsense/beta",
    messages=[{
        "role": "user",
        "content": "Generate a video: A golden retriever running through a sunlit meadow of wildflowers, slow motion, cinematic, shallow depth of field, warm afternoon light, camera tracking alongside the dog"
    }]
)

print(response.choices[0].message.content)
```

### Streaming

Set `"stream": true` for real-time Server-Sent Events (SSE) responses, or `"stream": false` to receive the complete result in a single response. Streaming is useful for showing progress in UIs; non-streaming is simpler for scripts and automation.

## Prompt Structure

```
Generate a video: [subject] + [action] + [environment] + [style] + [camera movement] + [lighting]
```

## Camera Movement Keywords

| Movement | Description | Use For |
|----------|-------------|---------|
| **Static shot** | Camera does not move | Portraits, product display |
| **Pan left/right** | Horizontal camera rotation | Revealing environments |
| **Tilt up/down** | Vertical camera rotation | Revealing tall structures |
| **Dolly in/out** | Camera moves toward/away | Dramatic focus shifts |
| **Tracking shot** | Camera follows subject | Action, movement sequences |
| **Crane shot** | Rising or descending movement | Epic establishing shots |
| **Orbit** | Camera circles the subject | 360-degree reveals, products |
| **Zoom in/out** | Focal length change | Emphasis, dramatic reveals |
| **Handheld** | Slightly shaky, natural feel | Documentary, intimate |
| **Drone shot** | Aerial perspective | Landscapes, establishing |

## Style Keywords

| Style | Keywords |
|-------|----------|
| **Cinematic** | cinematic, film look, anamorphic, 24fps, shallow DOF |
| **Photorealistic** | photorealistic, real footage, DSLR video, natural |
| **Animated** | 3D animation, cartoon, anime, cel animation |
| **Abstract** | abstract motion, fluid simulation, particle effects |
| **Vintage** | Super 8, VHS, film grain, retro, old footage |
| **Slow Motion** | slow motion, 120fps, bullet time |

## Examples

### Cinematic Nature Shot

```bash
curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a video: Aerial drone shot slowly rising over a misty mountain forest at sunrise, revealing snow-capped peaks in the distance, golden light breaking through clouds, cinematic 4K, smooth camera movement, epic landscape"
      }
    ],
    "stream": false
  }'
```

### Product Showcase

```bash
curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a video: A sleek black smartwatch rotating slowly on a reflective dark surface, studio lighting with soft highlights, camera orbiting 360 degrees, premium product commercial style, clean minimal background"
      }
    ],
    "stream": false
  }'
```

### Abstract Motion Graphics

```bash
curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a video: Abstract fluid simulation of metallic gold and deep blue liquids merging and swirling together, macro close-up, mesmerizing slow movement, reflective surfaces, dark background, satisfying visual"
      }
    ],
    "stream": false
  }'
```

### Character Animation

```bash
curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a video: A samurai warrior drawing a katana in a bamboo forest, cherry blossom petals falling in slow motion, dramatic side lighting, anime-inspired cinematic style, camera slowly pushing in, atmospheric fog"
      }
    ],
    "stream": false
  }'
```

## Prompt Tips

- **One action per clip**: AI video works best with a single, clear action rather than complex sequences
- **Specify camera movement**: Without it, you may get a static or randomly moving camera
- **Include lighting**: "golden hour," "studio lighting," "neon glow" dramatically shapes the mood
- **Keep duration realistic**: Current AI video models produce short clips (3-10 seconds)
- **Describe motion explicitly**: "walking slowly," "rotating 360 degrees," "rising upward"
- **Add "smooth" or "fluid"**: Helps avoid jerky or unnatural movement

## Video Planning Workflow

For longer sequences, generate individual shots and edit together:

```bash
# Shot 1: Establishing
curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [{"role": "user", "content": "Generate a video: Wide establishing shot of a futuristic city skyline at dusk, flying vehicles crossing the sky, neon lights turning on, drone camera slowly descending, cyberpunk, cinematic"}],
    "stream": false
  }'

# Shot 2: Medium shot
curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [{"role": "user", "content": "Generate a video: Medium shot of a woman in a futuristic jacket walking through a neon-lit street market, rain reflecting neon signs, camera tracking from the side, cyberpunk, cinematic"}],
    "stream": false
  }'

# Shot 3: Close-up
curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [{"role": "user", "content": "Generate a video: Close-up of a woman looking up in wonder, neon reflections on her face, rain drops, shallow depth of field, camera slowly pushing in, cyberpunk, cinematic"}],
    "stream": false
  }'
```

## Related Skills

- [Image to Video](../image-to-video/SKILL.md) — Animate still images
- [AI Cinematic Trailer](../ai-cinematic-trailer/SKILL.md) — Movie-style trailers
- [AI Video Loop](../ai-video-loop/SKILL.md) — Seamless looping videos
- [AI Slow Motion](../ai-slow-motion/SKILL.md) — Slow motion effects

## Related Models

- [wan-2-1-14b](../../../models/wan-2-1-14b/SKILL.md) — High quality video generation
- [hailuo-minimax-video-01](../../../models/hailuo-minimax-video-01/SKILL.md) — Fast video generation

## Documentation

- [each::sense Overview](https://docs.eachlabs.ai/sense/overview)
- [each::labs API](https://docs.eachlabs.ai)

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.