# Use the AI tools

Ask the AI to edit in chat, make or change a picture, and use free tools that cut silence, find highlights, reframe, remove backgrounds and cut to the beat.

Source: https://infinall.ai/docs/ai-editor/ai-tools
Last checked: 2026-10-26

## What it is

The AI Editor is the video editor in your browser at /ai-editor. You cut clips on a timeline, add text and captions, effects, transitions, colour grading and audio, and export a finished video. It works on a computer only: phones and tablets see a message to use a computer, and the screen must be at least 1,024 pixels wide. Its AI tools are the **AI Editor** chat, which makes edits from your description, and **AI Gen**, which makes a picture from a description. It also has free tools that run in your browser: Auto Cut Silence, Local Highlights, Beat-Synced Auto-Edit, Auto Reframe, Background Removal and Quick Actions. See [open the editor](/docs/ai-editor/open-the-editor).

## Who can use it

Everyone. The app does not check your plan for the chat or for the free tools, so Free and Pro see the same controls. On a team, a paid AI action needs the Editor role or higher. A Viewer cannot spend credits.

One plan rule exists, in the picture maker. One picture model is marked **(Pro)**. A Free account that picks it gets "This picture model needs a Pro plan." The other models are open to everyone.

## Cost

Mixed.

- **Chat** and **Picture** use credits. You pay for what the AI really used. The credits are held first and anything not used is given back. If the call fails you are not charged. See [what is a credit](/docs/credits/what-is-a-credit).
- **Picture** shows an estimate on its button, for example "Generate image" followed by "about N credits". See [price before you confirm](/docs/credits/price-before-you-confirm).
- **Chat** shows no price before you send. Instead, one chat request stops after 100 credits.
- Editing, saving, exporting and every free tool below never use credits.

## Where to find it

- **Chat:** in the strip named **Editor sections** at the far right, choose **AI Editor**. It opens in the same panel as **Media**.
- **Picture:** in **Editor sections**, choose **AI Gen**, then **Picture**. For a picture you already have, hover it in **Media** and choose **Edit with AI**.
- **Auto Cut Silence:** select a clip with sound, stay on the **Edit** view and open **Auto Cut Silence**.
- **Background Removal:** select a video clip, open the **Effects** view and open **Background Removal**.
- **Auto Reframe, Beat-Synced Auto-Edit, Local Highlights, Quick Actions:** select a clip, open the **Effects** view and look in its AI part.
- **Beat Sync:** see [audio](/docs/ai-editor/audio).

## Buttons and controls

![The AI Editor chat panel with the model menu, Past chats, AI settings, New chat, the suggestions and the message box.](/docs/screens/ai-editor/ai-tools.webp)

<!-- shot: ai-editor/ai-tools -->

The numbered rows are the chat panel in the screenshot.

| # | Name in the app | What it does | Plan |
| --- | --- | --- | --- |
| 1 | AI model | A menu to choose the AI that does your edits. **Auto** is first. A **Default** tag marks the one the editor starts with. The menu hover text is "Choose ai model". It is greyed out while the AI is working. | Both |
| 2 | Past chats | Opens the list of your earlier chats. Greyed out while the AI is working. | Both |
| 3 | AI settings | Opens a small menu with two switches: **Explain first** and **Don't ask before deleting**. A dot shows on the button when one is on. | Both |
| 4 | New chat | Starts an empty chat. Greyed out while the AI is working or when the chat is already empty. | Both |
| 5 | What should I change? | The heading of an empty chat. Under it: "Describe the edit in your own words. I make it on your timeline, and you can undo it." | Both |
| 6 | Try one of these | Suggestion buttons. Click one to send it. With clips on the timeline they are edits such as "Add captions to my video". On an empty timeline they are starters such as "Add calm background music". | Both |
| 7 | For example: make the intro shorter | The message box. Its name for screen readers is "What should the AI change?". | Both |
| 8 | Send | Sends your message. It is greyed out while the box is empty. | Both |
| | Stop | Replaces **Send** while the AI works. Ends the request. | Both |
| | Working | A small tag in the header while the AI works. | Both |
| | Undo | Shows in the header after the AI changed something. Its hover text is "Undo what the AI just did". | Both |

**Cards in the chat.**

| # | Name in the app | What it does | Plan |
| --- | --- | --- | --- |
| | Okay to do this? | A card that appears before the AI removes or deletes something. It says "You can undo it afterwards." | Both |
| | Yes, do it | Lets the AI do this one step. | Both |
| | Yes to all | Lets the AI do every step of this request without asking again. | Both |
| | No | Skips the step. The step shows **Skipped**. | Both |
| | Working, Done, Didn't work, Skipped | The status on each step card the AI shows while it edits. Click a card to see its details. | Both |
| | Close message | On an error card. Clears the error. | Both |

**Picture (AI Gen).**

| # | Name in the app | What it does | Plan |
| --- | --- | --- | --- |
| | Make with AI | The group on the **AI Gen** screen that holds **Voice-over** and **Picture**. | Both |
| | Picture | Opens the picture maker. Its hint reads "Describe a picture and AI draws it for you". | Both |
| | Generate image | The panel title when you make a new picture. | Both |
| | Model | A menu of five picture models. Its name is "Picture model". One is marked **(Pro)**. | Both. One model is Pro. |
| | Size | A menu with **Square (1:1)**, **Landscape (3:2)** and **Portrait (2:3)**. | Both |
| | Prompt | The box for your description. It shows a count against the limit. | Both |
| | Generate image (with the estimate after it) | The button. While it works it reads **Generating...**. | Both |
| | Generated result | The picture that appears under the button, with the line "Added to your media" and the model name. | Both |
| | All tools | The back link, to return from **Picture** to the **AI Gen** list. | Both |

**Free tools in the inspector.**

| # | Name in the app | What it does | Plan |
| --- | --- | --- | --- |
| | Auto Cut Silence | Finds quiet gaps in a clip's sound and cuts them out. Buttons: **Analyze**, then **Cut N**. | Both |
| | Beat-Synced Auto-Edit | Cuts your video to the beat of a music clip. Buttons: **Generate Auto-Edit**, then **Apply Auto-Edit**. | Both |
| | Local Highlights | Finds the most energetic moments in a clip. Buttons: **Find Highlights**, then **Apply N Highlights**. | Both |
| | Auto Reframe | Changes the project to a new shape. Button: **Analyze & Reframe**. Video clips only. | Both |
| | Background Removal | Hides or replaces the background behind a person. Video clips only. | Both |
| | Quick Actions | Three one click buttons: **Remove Background**, **Quick Dialogue Cleanup** and **Auto-Color**. | Both |

## How to ask the AI to edit

1. In **Editor sections**, choose **AI Editor**.
2. Open or start a project first. With no project open the panel says "Open or start a project first, then tell me what to change."
3. Optional: pick a model in **AI model**. **Auto** picks a lighter model for simple edits and a stronger one for bigger jobs.
4. Type what you want, for example "Trim 2 seconds off the end of the first clip". Press **Enter** or click **Send**. **Shift** and **Enter** start a new line.
5. Watch the step cards. If the AI wants to remove something, answer the **Okay to do this?** card.
6. When it is done, check the timeline. To take the change back, click **Undo** in the chat header.

If a request hits the limit it says "This request used 100 credits and stopped. Say continue to keep going." Type **continue** to carry on.

## How to see what the AI would do first

1. Click **AI settings**.
2. Turn on **Explain first**. The hint says "Tell me the plan without changing anything."
3. Send your message. The AI tells you its plan and changes nothing.
4. Turn the switch off again when you want it to make the edit.

## How to make a picture

1. In **Editor sections**, choose **AI Gen**, then **Picture**.
2. Pick a **Model** and a **Size**.
3. Type what you want in **Prompt**. Be specific about the subject and the light.
4. Check the estimate on the button. Click **Generate image**.
5. Wait. The panel says "Images usually take 20 to 60 seconds."
6. The picture is added to **Media**. Drag it onto the timeline.

## How to change a picture with AI

1. In **Media**, hover a photo and click **Edit with AI**.
2. In the **Edit image with AI** window, pick a **Model** and **Size**.
3. In **What should change?** describe the change, for example "replace the background with a sunset beach".
4. Click **Edit image**. The result is added to **Media** as a new file. Your original stays.

## How to cut silence

1. Select a video or audio clip and open **Auto Cut Silence**.
2. Move **Silence Threshold** and **Min Duration** if you need to. The note says "Lower values detect more silence".
3. Click **Analyze**. You see **Silent Sections Found** and **Total Silence**.
4. Click **Cut N**. You see "Silence Removed".
5. To undo every cut at once, press Ctrl+Z. On a Mac use Command+Z.

## How to find highlights

1. Select a clip with sound and open **Local Highlights**.
2. Set **Clips** (how many you want) and **Max duration** (in seconds).
3. Click **Find Highlights**. Each moment has a score, a title, a time range and a reason.
4. Click a moment to include or leave it out. Click **Preview highlight** to move the playhead there.
5. Click **Apply N Highlights**. The clip is split and the parts you did not pick are removed.

## How to cut video to the beat

Two tools do this. **Beat Sync** is in the **Edit** view when you select an audio clip. See [audio](/docs/ai-editor/audio).

**Beat-Synced Auto-Edit** works like this:

1. Put at least one music or sound clip and one video clip on the timeline. With no sound the panel says "Add an audio track to use auto-edit".
2. Select a clip and open **Beat-Synced Auto-Edit**.
3. Choose the sound in **Audio Source**.
4. Choose a **Cut Mode**: **beats**, **downbeats** or **segments**.
5. Set **Sensitivity** and **Min Clip Duration**.
6. Click **Generate Auto-Edit**. You see **Preview** with the number of **Cuts** and the **Duration**.
7. Click **Apply Auto-Edit**. The video row is replaced by the cut clips.

## How to reframe for another shape

1. Select a video clip and open **Auto Reframe**.
2. Click a platform under **Platform Presets**, or a shape under **Aspect Ratio**.
3. Click **Analyze & Reframe**. You see "Auto Reframe Applied" and the new size.

## How to remove or replace a background

1. Select a video clip, open the **Effects** view and open **Background Removal**.
2. Click **Off** to turn it **On**. The first time, the app loads its picture model, so allow a moment. You see "Background Removal Ready".
3. Under **Background Mode**, choose **Blur**, **Color**, **Image** or **Transparent**.
4. Adjust **Blur Amount**, pick a colour, or click **Choose Image** for your own picture.
5. Use **Edge Smoothing** and **Detection Threshold** to tidy the edge.

## Options explained

**Chat models.** The menu groups models by the company that makes them. **Auto** chooses for you: a short, simple request goes to a lighter model, and a long request, a long chat or a big job such as "build", "montage" or "captions for" goes to a stronger one. Every model is billed in credits for what it really used.

**AI settings.**

| Switch | What it does |
| --- | --- |
| Explain first | The AI tells you the plan and changes nothing. |
| Don't ask before deleting | The AI removes or replaces things without showing **Okay to do this?**. |

**What asks first.** The AI asks before steps that remove things: removing a clip, a row, a caption or a media file, and ripple delete. Looking at your timeline never asks.

**Picture models.** Five models are offered. Four are open to everyone. One is marked **(Pro)**. The editor picks the quality for you, so there is no quality menu. The model that answered is named under the result.

**Picture sizes.**

| Choice | Pixels |
| --- | --- |
| Square (1:1) | 1,024 x 1,024 |
| Landscape (3:2) | 1,536 x 1,024 |
| Portrait (2:3) | 1,024 x 1,536 |

**Auto Cut Silence.** **Silence Threshold** runs from -80 to -20 dB and starts at -40 dB. **Min Duration** runs from 0.1 to 2 seconds and starts at 0.5 second. **Pad Before** and **Pad After** each run from 0 to 2 seconds and start at 0.1 second. Padding leaves a little quiet at each side of a cut so it does not sound clipped.

**Cut Mode (Beat-Synced Auto-Edit).** **beats** cuts on every beat. **downbeats** cuts only on the strong first beat of each bar. **segments** cuts on every few beats, so you get longer clips. **Sensitivity** goes from 0 to 100 percent and starts at 50. A higher value keeps weaker beats, so you get more cuts. **Min Clip Duration** goes from 0.1 to 3 seconds and starts at 0.5.

**Local Highlights.** **Clips** goes from 1 to 20 and starts at 5. **Max duration** goes from 1 to 300 seconds and starts at 60. The score colour is green for 8 and up, yellow for 5 to 7, and grey below that. The work happens in your browser and finds loud, lively moments. It does not read what is said.

**Auto Reframe.**

| Group | Choices |
| --- | --- |
| Platform Presets | YouTube, TikTok, Instagram Reels, Instagram Feed, Instagram Stories, YouTube Shorts, Facebook, Twitter/X and LinkedIn. |
| Aspect Ratio | 16:9, 9:16, 1:1, 4:5, 4:3 and 21:9. |
| Sliders | Tracking Speed, Smoothing and Center Bias, and a **Follow Subject** switch. |

Choosing a platform also chooses its shape. In this version **Analyze & Reframe** resizes the whole project to the size you picked. The size line under the button shows the new width and height. The tracking sliders and **Follow Subject** do not change the result yet, so do not rely on them.

**Background Removal modes.** **Blur** blurs the background, with **Blur Amount** up to 50 pixels and a start value of 15. **Color** fills it with one colour from eight swatches or your own. **Image** uses a picture from your computer. **Transparent** leaves it empty. **Edge Smoothing** goes up to 10 and starts at 3. **Detection Threshold** is a percentage.

**Quick Actions.** **Remove Background** switches on a green screen key: it removes green from the picture. It is not the same as **Background Removal**, which finds a person. **Quick Dialogue Cleanup** turns on the **Speech Focus** noise cleanup. **Auto-Color** adds a small lift to saturation, contrast and brightness.

## Limits

- One chat request stops after 100 credits. Say continue to keep going.
- A picture you send for an AI edit can be up to 25 MB.
- A picture description can be up to 4,000 characters.
- Picture sizes are fixed at the three choices above.
- A single media file can be up to 500 MB when you bring it in.
- **Auto Cut Silence** only shows for a clip that has sound.
- **Auto Reframe**, **Background Removal** and **Remove Background** only show for a video clip.

## When something goes wrong

| Message in the app | What it means | What to do |
| --- | --- | --- |
| Request limit reached | This request used 100 credits and stopped. | Type **continue** to carry on. |
| Out of credits | You don't have enough credits for this AI request. Top up to keep using AI. Editing, saving and export stay free. | Click **Top up credits**. |
| Couldn't reach the AI service | Check your internet connection and try again. | Click **Try again**. |
| Rate limit reached | The AI service is busy. | Wait a moment and click **Try again**. |
| Conversation is too long | This model cannot fit the full conversation. Start a new chat and continue there. | Click **Start new chat**. |
| Model unavailable | The model you picked is not available right now. | Choose another model and try again. |
| AI request failed | The AI could not finish. The card shows the reason. | Click **Try again**, or change your wording. |
| The AI request was stopped. | You pressed **Stop**, or closed the tab. | Send the message again. |
| This picture model needs a Pro plan. | A Free account picked the model marked **(Pro)**. | Pick another model, or upgrade in Settings > Billing. |
| A prompt is required. | The picture box was empty. | Describe the picture first. |
| The prompt is too long. | The description is over the limit. | Shorten it. |
| Could not add the image to your media. | The picture was made but could not be saved. | Try again. |
| Image generation failed. | The picture could not be made and the app has no more detail. | Try again. |
| No Silence Detected | **Auto Cut Silence** found nothing at these settings. The note says "No silent sections found with current settings. Try lowering the threshold." | Move **Silence Threshold** toward -80. |
| Cut Failed | The cuts could not be made. | Click **Analyze** and try again. |
| Add an audio track to use auto-edit | **Beat-Synced Auto-Edit** has no sound clip to follow. | Add music first. |
| Audio file not loaded | The sound file is not ready in the browser. | Wait, or re-add the file. |
| Auto Reframe Failed | The reframe could not finish. | Try again. |
| Processing Failed | The background model could not start. The note says "Could not enable background removal". | Check your connection and click **On** again. |
| Could not clean up audio | **Quick Dialogue Cleanup** failed on this clip. | Open **Background Noise Removal** and pick a card there. |

## Questions

### Does the AI chat cost credits every time I send a message?

Yes. Each message uses credits for what the AI really used. A small edit costs little. A big job costs more. The chat does not show the price first, but a request stops after 100 credits, and unused credits are given back.

### Can I undo what the AI did?

Yes. Click **Undo** in the chat header right after the change. The same **Undo** in the top bar also works. If you made another edit by hand since then, the chat **Undo** button goes away so it cannot undo the wrong thing.

### Are my chats saved?

Yes, on this device. Click **Past chats** to search them, open one to keep going, or delete one. The list says "Saved on this device".

### Which tools are free?

Everything except the chat and **Picture** (and **Voice-over**, see [audio](/docs/ai-editor/audio)). Auto Cut Silence, Local Highlights, Beat-Synced Auto-Edit, Auto Reframe and Background Removal all run in your browser and use no credits.

### Does Free get fewer picture models?

Four of the five are open to Free. The one marked **(Pro)** needs a Pro plan.

### Can the AI make colour looks or captions?

Colour looks are in **AI Gen**, see [colour grading](/docs/ai-editor/colour-grading). Captions are in the **Captions** panel, see [text and captions](/docs/ai-editor/text-and-captions). Both are separate from the chat.