At Stream, we know moderation is critical for building safe, thriving communities. With millions of messages flowing through apps powered by Stream, companies need more than just static rules. They need flexibility, speed, and full control.
That's why we're excited to announce the release of AI Text Harm Detection (LLM). This new feature gives teams the ability to define, edit, and manage moderation harms directly from the Stream dashboard; no need to wait for us to configure changes.
Let's dive into what's new.
What's New with Editable LLM Prompts
- Full Flexibility: Define any type of harm in plain language, whether it's hate speech, self-harm, threats, or community-specific guidelines.
- Dashboard Control: Create, edit, and manage harms in real time without contacting Stream support. Your moderation rules are in your hands.
- Improved Detection: Built on LLM-powered detection, the system offers notable performance gains in accuracy and coverage.
- Default Availability: The feature is being rolled out to existing customers in phases and is automatically enabled for all new customers.
Why It Matters
Before, teams needed to rely on Stream to update harms or prompts. Now, moderation teams are in the driver's seat. You can quickly adapt to emerging risks, adjust for community needs, and ship updates without technical bottlenecks.
This shift means:
- Faster response to harmful content trends.
- Tailored moderation aligned with your unique community guidelines.
- Less dependency on backend updates or external approvals.
What's Next
This is just the start of our AI moderation journey. With flexible, LLM-powered text harm detection now in your hands, we'll continue expanding coverage and refining performance. As always, your feedback will guide the roadmap.
Get Started
- Start your free trial today.
- Reach out to our sales team to explore how Stream Moderation can work for your business.
- Have feedback? We'd love to hear it at moderation@getstream.io.
Thanks for building safer, more engaging communities with us.