We use advanced language models to analyze the context in which words are used. This allows us to accurately identify and handle profanity based on its context, ensuring that only inappropriate uses are flagged.
Yes, we support escalations to manual moderators for cases that require human review. This ensures that complex or ambiguous cases are handled appropriately.
Yes, we have regional endpoints to ensure low latency and compliance with regional data protection regulations.
Yes, we provide real-time query tracking to help you monitor and manage your content moderation in real-time.
Yes, you can access detailed moderation analytics through our dashboard. This includes insights into flagged content, moderation decisions, and trends over time.
Our platform is designed to handle high volumes of concurrent traffic, ensuring reliable performance even during peak times. Specific limits can be discussed based on your needs.
Our p99 latency is 350ms for text, and our p50 latency is 100ms, ensuring fast and efficient moderation.
Yes, we support adversarial spelling detection, which is one of our most critical features. This helps in identifying and moderating content that uses misspelled words to bypass filters.
We support the following languages:
Yes, our tool allows you to block specific topics, including competitor mentions, regulatory terms, cancellation requests, and password resets.
Our service utilizes advanced language models (LLMs) that provide context-aware moderation, which is far superior to simple keyword matching or regexes. This ensures more accurate and reliable content moderation.
The maximum character length for content that can be moderated by our service is 5,000 characters.
Yes, we offer the option to have a dedicated host for your services on your preferred cloud provider, ensuring better performance and security.
Content escalation requests are for content we consider suspicious. We may ask you to submit a second request to review such content to reduce false positives.
No, we do not provide a list of banned keywords as part of our service.
