> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://fastpix.com/docs/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://fastpix.com/docs/_mcp/server.

# Create a moderation rule

A rule is where you set your moderation policy: what Advanced Moderation looks for in the video and audio, how strict to be, and what to do when it finds a match. You encode a use case in a rule, such as a family-safe library, brand-safe ads, or screening user-generated uploads, then reuse it across jobs. For the bigger picture, see the [Advanced Moderation overview](/docs/in-video-ai/advanced-moderation).

You create and manage rules in the dashboard under **AI Agents > Moderation Agent > Rules**.

![FastPix Moderation Agent welcome screen](https://static.fastpix.com/fastpix-moderation-agent-welcome.png)

\


## Create a rule

1. In the dashboard, go to **AI Agents > Moderation Agent > Rules**, then click **Create rule**.
2. Under **Rule details**, enter a **Rule name** you will recognize when you pick it for a job, for example `Family-safe library`.

![Create rule screen with rule name and conditions](https://static.fastpix.com/fastpix-create-rule-empty.png)

\


3. Under **Conditions**, choose what to look for. Turn on a whole category, or expand it to pick individual items. Each condition has a threshold and an action.

![Create rule conditions expanded with video and audio categories](https://static.fastpix.com/fastpix-create-rule-conditions-expanded.png)

\


### What you can detect

FastPix moderates two kinds of content:

* **Video conditions:** Violence, Graphic violence, Nudity, Minors present, Drugs & alcohol, Smoking & tobacco, Sexual involving minors, Child abuse, Religious sensitive, Religious desecration.
* **Audio conditions:** Profanity, Harassment, Threats, Sexual involving minors, Child abuse, Religious sensitive, Religious desecration.

\


### Set the threshold

Each condition has a **threshold** from `0.1` to `1.0` that controls how sensitive the detection is. FastPix acts on a moment when its risk score reaches the threshold, so a **lower value flags more moments** (more sensitive) and a **higher value flags only higher-risk, clear-cut moments**. Tune it per condition to balance catching everything against false positives. You can set one value for a whole category, or a different value for each item.

\


### Choose an action

For each condition, choose what FastPix does when it finds a match:

* **Blur** (video): blurs the picture while the flagged moment is on screen.
* **Beep** (audio): covers the flagged words or sounds with a beep.
* **Mute** (audio): silences the audio while the flagged moment plays.
* **Cut** (video): removes the flagged moment from the video.
* **Flag**: marks the moment in the results without changing the video.

![Create rule with conditions, thresholds, and actions configured](https://static.fastpix.com/fastpix-create-rule-configured.png)

\


4. Click **Create rule**. Your rule appears in the **Rules** list, ready to use in a job.

![Moderation Rules list](https://static.fastpix.com/fastpix-moderation-rules-list.png)

\


> **NOTE:** You can create multiple rules for different needs. For example, one rule that blurs violence and mutes religious-sensitive audio, and another that beeps all profanity. You pick which rule to apply when you create a job.

![Create rule example mixing blur, flag, beep, and mute actions](https://static.fastpix.com/fastpix-create-rule-violence-religious.png)

\


#### Frequently asked questions

Content moderation detects and flags sensitive content so you can decide what to do with it. Advanced Moderation goes further: it acts on the flagged moments automatically, blurring the video, beeping or muting the audio, or cutting the moment out, and produces a moderated copy of the video alongside the original.

The threshold (0.1 to 1.0) sets how sensitive a condition is. FastPix acts on a moment when its risk score reaches the threshold, so a lower value flags more moments and a higher value flags only the most clear-cut ones.

Yes. A single rule can include video conditions (such as violence or nudity) and audio conditions (such as profanity or threats), each with its own threshold and action.

\


## What's next

* [Run a moderation job](/docs/in-video-ai/run-a-moderation-job) to apply this rule to your videos.
* [Review moderated media](/docs/in-video-ai/review-moderated-media) to see what was flagged and compare the moderated copy with your original.