Video summary

Runway AI - Tutorial for Beginners in 17 MINUTES ! [ FULL GUIDE ]

Main summary

Key takeaways

Technology

Summary of the tutorial (Runway AI for beginners)

The video demonstrates Runway, a browser-based AI creative toolkit for generating and editing images and videos. It walks through the UI layout, core generation workflows, model choices, and practical techniques such as references, keyframes, sketch-based control, and audio/lip-sync.

1) Runway interface & where tools are

  • Access Runway in a web browser (no installation).
  • After logging in:
    • Sessions: projects in progress
    • Assets: uploaded and generated files
    • Create / All Tools: lists AI generation features
  • Start generation via:
    • Start a Session (workspace)
    • Chat Mode (chatbot interface)

The tutorial distinguishes between:

  • Chat Mode: simpler, beginner-friendly generation and follow-up suggestions
  • Tool Mode: more advanced controls (manual inputs, model settings, etc.)

2) Chat Mode generation (images and video)

  • Request content by prompt, such as:
    • “blonde horse running in the desert”
    • or a request for a video
  • You can upload reference images to guide generation.
  • Prompt guidance:
    • Use direct, clear prompts
    • Avoid “loops” or “negative phrases”
    • Include specific subject/action details
  • Runway uses Gen-4 (highlighted as the newest model in the tutorial).
  • After generation, Runway provides follow-up options (e.g., restyle into a cartoon).

3) Tool Mode video generation with Gen-4 Turbo

In Tool Mode, you explicitly choose:

  • Generate Image or Video
  • AI model (Gen-4 Turbo / Gen-3 variants)

Model availability notes:

  • Gen-4 and Gen-3 are supported
  • Gen-1/Gen-2 are not available in the newest version

Gen-4 Turbo positioning

  • Strong for live actions, VFX, motion clips
  • Good for animating simple images
  • Uses a reference image (from Assets or created via Create Image)

Video prompt structure (demonstrated)

  • Main subject + movement
  • Camera motion (e.g., track the horse)
  • Scene/background (e.g., desert static; dust rises)
  • Style (cinematic, realistic, cartoon, 3D, etc.)

Video controls & output

  • Aspect ratio
  • Length (example: 5 seconds)

Output details:

  • Preview player shown
  • Download available in 720p at 24 fps
  • 4K upscaling costs credits

Credits & plans

  • Free plan starts with 125 credits
  • Credits depend on image/video type and selected model
  • Paid plans can unlock more credits, watermark removal, and 4K upscaling

Copyright claim (as stated)

  • Generated media copyright is fully owned by the user
  • Can be used freely, including commercial purposes (per the tutorial)

4) Gen-4 (stronger) and Aleph video editing

  • If Gen-4 Turbo isn’t sufficient, switch to Gen-4 (paid plans only).

Claimed advantages of Gen-4:

  • Better character consistency
  • Stronger alignment with reference images

Aleph editing (Runway Aleph)

  • Edit existing video up to the first 5 seconds (per the tutorial).
  • Workflow:
    • Choose Edit with Aleph
    • Describe changes using simple commands like Add / Remove / Change / Replace
    • Optionally include a reference image to guide edits
  • Example: replacing the horse with a fire horse

5) Gen-3 Alpha Turbo for experimentation + advanced controls

The tutorial covers Gen-3 Alpha Turbo:

  • Used to test ideas while consuming fewer credits than Gen-4 Turbo

Key feature:

  • Text Prompt with reference support (image/video as reference)
  • Video generation still at 720p

Advanced controls shown:

  • Extend: add up to 8 more seconds using a new prompt
  • Keyframes:
    • Control motion over time
    • Define one keyframe image and optionally add more (e.g., start/middle/end)
    • Tip: if frames look too different, generation may jump—use more similar keyframes for smoother motion
  • Camera Control:
    • Apply camera movement (panning/tilting) across the whole clip
    • Can also be included in the text prompt
  • Expand:
    • Change aspect ratio using Expand Video (e.g., portrait → landscape)

6) Gen-3 Alpha (text-only option; paid plan)

Gen-3 Alpha:

  • More precise, uses fewer credits than Gen-4
  • Paid plans only
  • Unlike Gen-3 Alpha Turbo, it can generate video using text prompt only
  • Reference image/video optional for more detail
  • Uses detailed prompts especially when not using reference images

Reference integration:

  • Reference can be inserted as the first or last keyframe
  • Useful for animating a static image

7) Image generation with Gen-4 + references using “@” labels

  • Switch to the Image tab to create images using the Gen-4 image model.

Prompt elements:

  • Subject + actions (if applicable)
  • Background
  • Colors
  • Style

If no reference image is used:

  • Choose from style templates (e.g., “Vivid”, “High Contrast”)

Image generation controls:

  • Aspect ratio
  • Number of variations
  • Resolution: 720p or 1080p

Reference-based generation

  • Upload multiple reference images
  • Assign each reference a name
  • Use @Name in the prompt to bind concepts (e.g., @Lion, @JaguarPattern, @Vulcan)

If results don’t match:

  • Use Generate Prompt to improve prompting based on the references

Additional tutorial emphasis:

  • A generated image can become a reference for subsequent video generation
  • Helps achieve consistent characters (same fur/identity details)

8) Upscaling workflow for images (and key limitation for video)

After generating an image:

  • Download the image
  • Use Upscale Image to upscale to HD / 2K / 4K

Important limitation:

  • Video generations remain at 720p, even if you use a 4K reference image.

9) Character/guided video via Act Two (gesture/acting transfer)

In the video session, Act Two enables gesture-based character motion:

  • Provide a reference video (record or upload)
  • Upload the character (image or video reference)

Options include:

  • Aspect ratio
  • Expressiveness
  • Whether to transfer gestures

Output length:

  • Generated video lasts as long as the reference video

For facial expressions only:

  • Use Gen-3 Alpha Turbo with Act One (noted as having future depreciation)

The tutorial points viewers to a previous Runway tutorial for Act One details.

10) Adding voice via Generative Audio + Lip Sync Video

To add audio:

  • Go to Generative Audio
  • Use Lip Sync Video

Workflow:

  • Import a character video (templates possible)
  • System analyzes faces (primarily human faces)
  • Provide an audio file or type text to be spoken
  • Choose a voice using the microphone button

The tutorial shows example output speech.

11) Sketch feature for precise image/video control

The tutorial demonstrates:

  • Creating a sketch reference:
    • Add a named image reference (e.g., a tree)
    • Use brush/sketch tools to draw shapes or mark areas
    • Undo with Ctrl/Cmd+Z
  • Use Export Sketch:
    • Combine sketch reference with a prompt to generate changes

Additional sketch controls:

  • Apply edits to specific regions using a standout color (example: pink)
  • Add arrow + label to clarify what the AI should do (e.g., pull down a branch)

Sketch can also create camera shot concepts.

Main speakers / sources

  • Primary source: The tutorial’s narrator/host (instructional walkthrough)
  • Tool/Platform source: Runway AI (official interface behavior as demonstrated)
  • Transcript source: TurboScribe.ai (auto-transcription for subtitles)

Original video