- XF 兼容性
- 2.3.x
- 摘要
- Relevio, a unique context-aware moderation engine for XenForo, reduces your moderation workload by up to zero. It uses advanced AI for rule-driven semantic checks before content is posted, ensuring only quality posts reach your server. Multimodal coherence and cryptographically verifiable decisions enhance security and user trust.
This is the XenForo add-on for Relevio, a unique context‑aware, rule‑driven moderation engine that enables autonomous forums by reducing moderation workload to or close to zero.
How does it work?
Relevio is a new class of moderation engine employing Rule‑Driven Semantic Moderation, which uses the full power of the most advanced large language models (LLM). This is very different from the established method of Category‑Based Classifier Moderation.
Moderation before submission
Uniquely, Relevio is invoked from the user interface before content ever reaches your server. This means your forum only ever receives quality checked content. This is the opposite of CBCM moderators that work on content after it reaches your server, by automatically flagging it or moving it to your moderation queue. Such systems can actually increase your moderation workload. Relevio is designed to reduce your workload in absolute terms, by preventing bad posts from being created in the first place.
Rule‑driven, not category‑driven
With RDSM you define the rules in plain English. Relevio applies them consistently, using AI only as a semantic interpreter. There is no drift, no unpredictability and no black‑box behaviour. There are no fixed categories and no percentage scores. Each rule either passes or fails.
Context‑aware by design
Every check includes context (the related story and prior reply chain). Relevio understands where a new post sits in the conversation, so it can enforce relevance and keep threads coherent.
Multimodal coherence
Relevio checks that images, attachments and text actually belong together, evaluating them as a single semantic unit. This prevents users from pairing misleading images or attachments with unrelated narratives and keeps your platform coherent, trustworthy, and harder to exploit.
Cryptographically verifiable decisions
Every approved submission is signed with an HMAC that binds the content, images, ruleset, and context together with your private key. The XenForo backend only accepts trusted, tamper‑proof posts. It is this trust that allows the moderation to take place in the user interface.
Instant feedback for users
Relevio presents users with clear, structured explanations and actionable suggestions before they submit. This reduces the irritation and distrust people feel when their content is rejected by a black‑box system. Instead of “your post was blocked,” users get transparent guidance that helps them fix issues quickly and stay engaged.
Multiple rulesets
Relevio allows you to define multiple rulesets, with different rules for new posts, replies, and across different forums / nodes. This allows you to keep individual forums focussed on particular topics. Rulesets are managed from your relevio.ai account.
Fast
Moderation time is 1-2 seconds for straight text with no rule violations, slightly longer for failures and around 10-15 seconds for multi-modal content (images, etc) depending on complexity.
Fail safe
The Relevio XenForo add-on has a fail safe design. If the main Relevio service is unavailable or faults, then posts are automatically sent to the normal moderation queue. This ensures that your system is never blocked from receiving new posts.
Internally the Relevio.ai moderation engine uses a redundant architecture, which means it automatically falls back on backup LLMs should our primary LLM fail.
Easy to configure
All you need is a relevio.ai account. Simply install the add-on, enter your account id and private key in the add-on settings and you are ready to go.
Bypass permission
Grant user groups permission to bypass Relevio moderation and allow your trusted users, administrators and moderators to post content without constraint.
Auditing and statistics
Requests to Relevio moderation are fully audited and a range of statistics reports are available for analysing usage, costs, rule failures, etc.
Costs
30 free moderations every month
The XenForo add-on requires a relevio.ai account and an API private key to function.
For the XenForo add-on, you should use the Relevio API-only pay-as-you go plan. Charging is based on processing units allowing the cost of moderating to scale with the size of post and whether there are images involved.
Processing units are calculated as follows:
Up to 1000 characters of input text: 1 unit
Each additional 1000 characters of text: +1 unit
Each image up to 1024×1024 pixels: +1 unit
Each attachment: +1 unit
Check our website for current unit pricing
A call to the moderation API is charged whether it returns pass or fail. A minimum of 1 processing unit applies to all calls to the moderation API.
The Pay-as-you go plan currently includes 30 free processing units every month to aid testing and evaluation.
Requirements
This version has been tested on XenForo 2.3.x and XenForo 2.2.x
How does it work?
Relevio is a new class of moderation engine employing Rule‑Driven Semantic Moderation, which uses the full power of the most advanced large language models (LLM). This is very different from the established method of Category‑Based Classifier Moderation.
Moderation before submission
Uniquely, Relevio is invoked from the user interface before content ever reaches your server. This means your forum only ever receives quality checked content. This is the opposite of CBCM moderators that work on content after it reaches your server, by automatically flagging it or moving it to your moderation queue. Such systems can actually increase your moderation workload. Relevio is designed to reduce your workload in absolute terms, by preventing bad posts from being created in the first place.
Rule‑driven, not category‑driven
With RDSM you define the rules in plain English. Relevio applies them consistently, using AI only as a semantic interpreter. There is no drift, no unpredictability and no black‑box behaviour. There are no fixed categories and no percentage scores. Each rule either passes or fails.
Context‑aware by design
Every check includes context (the related story and prior reply chain). Relevio understands where a new post sits in the conversation, so it can enforce relevance and keep threads coherent.
Multimodal coherence
Relevio checks that images, attachments and text actually belong together, evaluating them as a single semantic unit. This prevents users from pairing misleading images or attachments with unrelated narratives and keeps your platform coherent, trustworthy, and harder to exploit.
Cryptographically verifiable decisions
Every approved submission is signed with an HMAC that binds the content, images, ruleset, and context together with your private key. The XenForo backend only accepts trusted, tamper‑proof posts. It is this trust that allows the moderation to take place in the user interface.
Instant feedback for users
Relevio presents users with clear, structured explanations and actionable suggestions before they submit. This reduces the irritation and distrust people feel when their content is rejected by a black‑box system. Instead of “your post was blocked,” users get transparent guidance that helps them fix issues quickly and stay engaged.
Multiple rulesets
Relevio allows you to define multiple rulesets, with different rules for new posts, replies, and across different forums / nodes. This allows you to keep individual forums focussed on particular topics. Rulesets are managed from your relevio.ai account.
Fast
Moderation time is 1-2 seconds for straight text with no rule violations, slightly longer for failures and around 10-15 seconds for multi-modal content (images, etc) depending on complexity.
Fail safe
The Relevio XenForo add-on has a fail safe design. If the main Relevio service is unavailable or faults, then posts are automatically sent to the normal moderation queue. This ensures that your system is never blocked from receiving new posts.
Internally the Relevio.ai moderation engine uses a redundant architecture, which means it automatically falls back on backup LLMs should our primary LLM fail.
Easy to configure
All you need is a relevio.ai account. Simply install the add-on, enter your account id and private key in the add-on settings and you are ready to go.
Bypass permission
Grant user groups permission to bypass Relevio moderation and allow your trusted users, administrators and moderators to post content without constraint.
Auditing and statistics
Requests to Relevio moderation are fully audited and a range of statistics reports are available for analysing usage, costs, rule failures, etc.
Costs
30 free moderations every month
The XenForo add-on requires a relevio.ai account and an API private key to function.
For the XenForo add-on, you should use the Relevio API-only pay-as-you go plan. Charging is based on processing units allowing the cost of moderating to scale with the size of post and whether there are images involved.
Processing units are calculated as follows:
Up to 1000 characters of input text: 1 unit
Each additional 1000 characters of text: +1 unit
Each image up to 1024×1024 pixels: +1 unit
Each attachment: +1 unit
Check our website for current unit pricing
A call to the moderation API is charged whether it returns pass or fail. A minimum of 1 processing unit applies to all calls to the moderation API.
The Pay-as-you go plan currently includes 30 free processing units every month to aid testing and evaluation.
Requirements
This version has been tested on XenForo 2.3.x and XenForo 2.2.x
![FS: AI Auto-Moderation [LITE]](/data/resource_icons/27/27765.jpg?1779343865){kind=icon}
