--- source_url: https://www.pubnub.com/docs/message-processing/moderation/overview title: Moderation updated_at: 2026-09-30T07:20:08.000Z --- # Moderation ## Documentation index To discover more PubNub resources: 1. Fetch [PubNub's llms.txt](https://www.pubnub.com/llms-full.txt) for a list of available pages in Markdown format. 2. Identify relevant URLs from that index. 3. Fetch the target pages. Do not assume a path exists, always check the index first. PubNub provides two built-in tools for managing message content on your channels, both configured in the [Admin Portal](https://admin.pubnub.com/) through BizOps Workspace: * **Channel Monitor**: a manual moderation tool that gives moderators a live view of conversations they can act on in real time. * **Auto Moderation**: automated content filtering that runs before messages are published. The core distinction is timing. Auto Moderation runs before a message is published: a blocked message is never published, while a reported or word-masked message is still delivered to subscribers (see [Auto Moderation](#auto-moderation) below for what each action delivers). Channel Monitor runs after publication, giving moderators a live view of conversations they can act on in real time. | | Channel Monitor | Auto Moderation | | --- | --- | --- | | When it runs | After publish (reactive) | Before publish (proactive) | | How it works | Manual review by a moderator | Automated policy and AI | | Best for | Live supervision, responding to reports | Scalable filtering at volume | | Engineering effort | None | None (policy editor) | | Availability | Paid plans | Beta. Paid plans. | ## Channel Monitor Channel Monitor is a web-based tool in BizOps Workspace that gives moderators a live view of active conversations. Messages are delivered to subscribers first. Moderators then review and act. A moderator can watch up to five channels at the same time. From the monitor view, a moderator can: * Edit or delete individual messages. * Mute a user from sending to a channel. * Ban a user from a channel. * Highlight messages that contain specific keywords using word lists. * Review messages that users have flagged as inappropriate. * Review messages that Auto Moderation has blocked, reported, or word-masked, under **Review moderated messages**. Blocked messages appear only there, not in the chat view, because they never reach the channel. Word lists in Channel Monitor highlight flagged words in messages for moderators. Channel Monitor shares word lists with Auto Moderation, so a list created in either module is available in the other. Channel Monitor is available on paid plans. Access it in the Admin Portal under **BizOps Workspace > Channel Monitor**. ## Auto Moderation [Auto Moderation](https://www.pubnub.com/docs/message-processing/moderation/configure-auto-moderation.md) filters message content before delivery using three independent features. Each feature runs as a Before Publish Function on the channels you select. What subscribers receive depends on the action a feature takes, not on a single blanket guarantee: * A **blocked** message is never published. The publisher gets an error instead. Auto Moderation keeps a record of every moderated message, including blocked ones, so moderators can review them in Channel Monitor. * A **reported** message is published unchanged. It's flagged in Channel Monitor for manual review, not withheld from subscribers. * A **word-masked** message is published with the configured words replaced by asterisks. * If the moderation check itself fails (for example, the moderation service is unreachable), Auto Moderation delivers the message unmoderated rather than blocking it. Failing open is the intended default, so a moderation service problem doesn't stop messages across your channels. To fail closed instead, deploy with **Generate code for Functions** and change the generated code, as described in [Configure Auto Moderation](https://www.pubnub.com/docs/message-processing/moderation/configure-auto-moderation.md#generate-code-for-functions). Auto Moderation is in Beta. It runs on [Functions](https://www.pubnub.com/docs/message-processing/serverless/overview.md) and is available on paid plans. Auto Moderation supports three features, which you can enable independently or together: * **Spam detection.** PubNub's AI model analyzes English-language messages. Messages identified as promotional spam or scam are either blocked before publish or reported to Channel Monitor for manual review. * **Profanity filter.** The Auto Moderation profanity filter analyzes messages across 15 supported languages against configurable toxicity categories (toxicity, harassment, hate speech, sexual content, spam, profanity, self-harm, doxxing, scam, impersonation, raiding, and political abuse). You set a detection threshold per category. Messages that exceed a threshold are blocked or reported. * **Word masking.** Messages containing words from your configured word list have those words replaced with asterisks before delivery. Auto Moderation and Channel Monitor share word lists. You configure Auto Moderation through a policy editor in BizOps Workspace. A policy sets the channels to protect, the features to enable, and the path to message content in your JSON payload. Different channels can have different policies. Auto Moderation works with both Core SDK and Chat SDK apps. It operates at the message level, so the SDK used to publish does not affect how filtering is applied. Two deployment options are available when you activate a policy: * **PubNub managed.** PubNub deploys and manages the Before Publish Function automatically. You can toggle it on and off without writing code. * **Generate code for Functions.** PubNub generates the moderation code for you to add to an existing Before Publish Function on the channel. Use this option when a Before Publish Function already runs on the channel, since only one can run per channel. :::note Auto Moderation availability Auto Moderation is in Beta. Contact [PubNub Support](mailto:support@pubnub.com) or [Sales](https://www.pubnub.com/company/contact-sales/) to enable it. ::: ## Using both together Channel Monitor and Auto Moderation complement each other. A common pattern is to use Auto Moderation to block or flag high-confidence violations automatically, then use Channel Monitor to review flagged messages, handle edge cases, and respond to user reports. Last updated at: 2026-09-30T07:20:08.000Z