I tried Google's free AI video tools: 3 that are actually worth using

There are now dozens of AI video tools promising studio-quality output, but most free options are either too limited or too hard to set up. Google, however, has quietly released several powerful AI tools you can use right in your browser—no credit card, no complex installs.

This guide walks through three of Google’s best free options for video and creative work: Google Vids, Google Flow, and Pomelli. You’ll see what each one does well, how to use them step by step, and when to pick one over the other.

1. Google Vids: the fastest way to start with AI video

Google Vids is the easiest entry point if you just want to generate and edit AI videos without touching complex settings or code. Everything runs in your browser and you can go from idea to finished clip in a few minutes.

Getting started with Google Vids

To open Google Vids, simply go to vids.new in your browser. That instantly creates a new project.

First, you choose your video orientation:

• Landscape
• Portrait
• Square

Landscape generally gives you the most flexibility, especially for YouTube-style content, so it’s a safe default.

Next, you’ll see a grid of ways to start a project, including:

• Record yourself
• Upload your own files
• Turn slides into video
• Create AI videos (text-to-video)
• AI avatar
• Editing tools

The three options that matter most for AI video creation are Create AI videos, AI avatar, and the Edit feature.

Create AI videos from a simple prompt

To generate a video from text, click Create AI videos. A panel opens on the right where you can:

• Paste your text prompt
• Optionally add reference images to guide the look

After you hit Generate, Vids creates a short AI video based on your description. For a free browser tool, the quality is surprisingly strong—especially considering that Google’s underlying models (like Omni Flash for video and Nano Banana for images) rank near the top on community leaderboards such as OpenArt Arena.

Results won’t be perfect: you might see minor issues like slightly odd motion or timing. But for quick concept clips, B-roll, or social posts, it’s more than usable.

Edit vs. recreate: the key setting most people miss

Once your video is generated, you’ll see three options underneath:

• Extend – make the clip longer
• Edit – change details in the existing video
• Recreate – generate a brand new video from scratch

This is where many people go wrong. If you like your clip but want to tweak one detail—like changing a shirt color or time of day—use Edit, not Recreate.

With Edit, you simply describe the change in plain language (for example: “Make his shirt red and change the scene to nighttime”). The tool keeps everything else consistent while applying your requested change.

Recreate, on the other hand, throws away the old clip and makes a new one. Use it only when you want a completely different take.

Use AI avatars to generate talking-head videos

Google Vids also includes built-in AI avatars you can use to generate talking-head videos from a script—great for explainer videos, social content, or automated training clips.

To use an avatar:

1. Go back to the main Vids screen and choose AI avatar.
2. Browse the preset avatars and hover to preview their voices.
3. Select one (for example, an avatar named Eric).
4. Paste in the script you want them to read.
5. Click Play preview to hear the voice.
6. Once you’re happy, hit Generate.

If your project has multiple scenes, you can choose whether to generate only the current scene or all scenes at once. Combined with the edit tools, you can build full videos that mix AI avatars, AI-generated B-roll, and your own footage—without paying anything.

2. Google Flow: advanced AI filmmaking with an agent

Google Vids is great for quick, simple videos. When you want more control—cinematic shots, multi-angle scenes, consistent characters, and smarter prompting—Google Flow is the better choice.

Flow is built around an AI agent that helps you plan and generate images and videos. Instead of writing perfect prompts yourself, you describe what you want in natural language and let the agent handle the technical details.

Setting up your first project in Google Flow

After opening Google Flow, start by creating a new project:

1. Click New project. Everything you generate will live inside this project space.
2. In the prompt box, click Agent and then the expand icon to open the full side panel.
3. If you ever prefer manual control, you can click the model name in the prompt box and set all parameters yourself—but the agent is usually easier.

Before generating anything, it’s worth configuring a few defaults:

1. Click the Settings icon at the bottom of the prompt box.
2. Set Confirm before generating to Always. This makes Flow show you what it plans to do and ask for approval before spending compute.
3. For Image defaults, choose 16:9 with the Nano Banana 2 model.
4. For Video defaults, choose 16:9 with Omni Flash.

Generating high-quality images with the AI agent

To test the setup, switch the image model to Nano Banana Pro (Google’s top-tier image model) and ask the agent for an image in plain language. For example, you might request:

A wide cinematic image of a woman who has just landed a small plane in a salt desert, standing by the aircraft, looking thoughtful.

When you send the request:

• The agent explains what it’s about to do.
• If you click Show thinking, you can see how it plans the prompt and settings.
• After a moment, the generated image appears directly in the chat.

The results are usually detailed and realistic, with strong lighting, composition, and storytelling cues—ideal for concept art, thumbnails, or reference frames for later video generation.

Choosing the right video model: VO 3.1 vs Omni Flash

Flow gives you access to two main video models, and choosing the right one matters a lot for your final result:

• VO 3.1 – higher-end cinematic quality, but limited to 8 seconds and no built-in camera cuts setting.
• Omni Flash – supports longer clips (e.g., 10 seconds or more) and is more flexible for varied content.

If you want short, highly cinematic shots and don’t mind the 8-second limit, VO 3.1 can look great. If you need longer scenes or more flexibility, use Omni Flash.

How to do multi-shot scenes with timestamp prompting

VO 3.1 doesn’t have a separate setting for camera cuts, so you need to write them into your prompt using timestamps. Google calls this timestamp prompting.

For example, imagine a fish market scene. Your prompt might look like this (simplified):

• 0–2s: Wide shot of a busy fish market at dawn, people setting up stalls.
• 2–5s: Cut to close-up of a fishmonger arranging fresh fish on ice.
• 5–8s: Cut to low-angle shot of customers walking past stalls, warm light filtering through.

At the end of the prompt, you specify what stays consistent across all shots: the same market, lighting, and characters. When you send this to the agent with VO 3.1 selected:

• The agent summarizes the job and asks for approval (thanks to the Confirm before generating setting).
• Once you approve, it generates a single 8-second clip with the specified cuts and actions.

The result can look surprisingly cinematic, with well-framed angles and smooth transitions—though character realism may still lag slightly behind dedicated high-end tools.

Letting the agent pick the best model for your request

One of the smartest parts of Flow is that the agent understands each model’s limits. For example, if you ask for:

A 10-second video of a small orange fox in a hand-drawn storybook style.

The agent knows VO 3.1 only supports up to 8 seconds. It will respond with options like:

• Switch to Omni Flash to keep 10 seconds, or
• Shorten the video to 8 seconds and stay on VO 3.1.

If you choose to switch to Omni Flash, Flow will generate the full 10-second storybook-style animation, keeping the fox consistent from start to finish.

This model-aware behavior makes it much easier to focus on creative direction instead of technical constraints—similar to how the best tools in roundups like the only 3 free AI video generators most creators actually need try to hide complexity behind smart defaults.

Solving character consistency with Flow’s character system

When you start stitching together multiple scenes into a longer story, a big problem appears: character consistency. Many AI video tools subtly change a character’s face, age, or clothing from shot to shot.

Google Flow tackles this with a built-in Characters feature.

Creating a reusable character

To create a character in Flow:

1. Click Characters in the left sidebar.
2. Either start from one of the sample characters or create a new one from scratch.
3. Write a full description of your character (appearance, age, clothing, vibe, etc.).
4. Optionally upload reference images, or rely purely on text for a unique look.
5. Click Generate to get an initial portrait.
6. If needed, generate additional angles by clicking the body icon.
7. Give your character a name (for example, Halver).
8. Click Select a voice to assign a voice preset or upload a recording to clone a real voice.
9. Fill in the Character info field with personality traits and behavior (for example: serious, reserved, speaks thoughtfully, values the sea and family).
10. Click Done to save the character to your project.

Using characters in your scenes

Back in the agent chat, you can now treat your character like a real actor:

1. Type @ in the prompt box to open the picker.
2. Select your saved character (e.g., Halver).
3. Write your scene prompt as if you’re directing them, including any dialogue.

For example, you might ask Omni Flash to create a 10-second scene of Halver on a fishing boat at sunset, looking out over the water and saying a reflective line about the sea.

When the clip is generated, the character’s look, voice, and personality stay consistent across shots. This is crucial if you want to build episodic content, brand mascots, or recurring protagonists without constant re-tuning.

If you’re exploring lots of tools and workflows, this kind of character system is one of the reasons Google Flow stands out compared to many other options in the "I tried every free AI video generator" style reviews, such as this breakdown of free AI video generators in 2026.

3. Pomelli: on-brand marketing content from a single URL

While Vids and Flow focus on video and visual storytelling, Google’s Pomelli tool is aimed squarely at marketers, founders, and designers who need on-brand creative assets across web, images, and campaigns.

Pomelli works differently from typical prompt-based tools. Instead of starting with text, you start with a website.

Turn any website into a reusable “business DNA”

When you open Pomelli, it asks for just one thing: a URL. You can enter your own site or a brand you admire. Pomelli then analyzes that website and builds what it calls your business DNA.

This DNA includes:

• Typefaces and fonts
• Color palette
• Brand values
• Tone of voice

Everything is editable, so you can tweak colors, copy, or values at any time. If you don’t have a website yet, you can instead upload up to 20 files (like PDFs, images, and docs), and Pomelli will build the DNA from those.

Generate on-brand websites in a few clicks

Once your DNA is ready, you can ask Pomelli to generate website designs that match your brand style. It creates multiple variants using your colors, fonts, and tone.

For example, if you feed it a brand like Oatly, Pomelli can output mock websites that mimic the same playful typography, button styles, and color schemes—while still giving you original layouts and copy.

Auto-generate a full brand book

Pomelli can also create a complete brand book from your business DNA in just two steps:

1. Pick a cover design from several options.
2. Confirm up to seven images that will be used throughout the book.

Instead of using random mockups, Pomelli can pull real product photos from the source website. The end result is a polished, multi-page brand book that documents your look, voice, and core messaging.

This is extremely useful when you want to brief other AI tools or human collaborators. You can simply paste sections of the brand book into a new chat instead of re-explaining your style and positioning every time.

Run AI “photoshoots” and turn them into videos

One of Pomelli’s standout features—something even many paid tools lack—is its Photoshoot mode.

Here’s how it works:

1. Click Photoshoot in the interface.
2. Choose a specific product (for example, a particular flavor of oat drink).
3. Select a set of templates or style inspirations (e.g., lifestyle shots, studio product shots).
4. Set the aspect ratio, such as Story 9:16 for vertical content.
5. Click Generate.

Pomelli then outputs multiple high-quality product visuals in different contexts—perfect for ads, social posts, and landing pages. The results often look like they came from a professional shoot, despite requiring almost no effort.

On top of that, you can click Animate on any image to automatically turn it into a short video. This makes it incredibly easy to go from static creative to motion assets for Reels, Stories, and ads.

Create full marketing campaigns from one brief

Pomelli doesn’t stop at visuals. You can also generate complete, on-brand marketing campaigns.

To do this:

1. Click Create campaign.
2. Describe your campaign in a single line (for example, “A playful summer campaign promoting our new oat drink flavor”).
3. Pomelli generates an editable brief with:

• Campaign title
• Description
• Goal
• Offer

You can adjust any of these before confirming. After that, Pomelli produces a full campaign concept with copy and creative ideas that match your brand voice.

The copy isn’t generic filler—it often includes clever, on-tone lines that feel close to what a human copywriter might produce for a modern consumer brand.

When to use each Google tool

All three tools are free, but they shine in different scenarios:

• Google Vids – Best for beginners and quick projects. Use it when you want fast AI videos, simple edits, and plug-and-play AI avatars without worrying about technical settings.
• Google Flow – Best for creators and filmmakers who want more control. Use it for cinematic shots, multi-angle scenes, and consistent characters across multiple videos.
• Pomelli – Best for marketers, founders, and designers. Use it to build brand DNA from a website, generate on-brand sites and brand books, run AI photoshoots, and launch full campaigns.

Used together, they cover a huge range of workflows—from concept art and character-driven stories to polished marketing campaigns and product promos—all without paying for a subscription.

Final thoughts

Most free AI video tools either lock the best features behind paywalls or require complicated setups. Google’s current stack is different: Vids, Flow, and Pomelli give you a surprisingly powerful toolkit for video, branding, and campaigns, all running in the browser.

If you’re experimenting with AI for content creation, these three are well worth adding to your workflow before you spend money on premium tools.