Spark Update, Safety CRM, and Landing Page
Summary
Mr. Anoop's server shares a two-week product update covering insights from CMX, a revamped landing page, and a new safety and moderation CRM with granular strike controls. The Spark backend now includes a web widget and a model switch to Google Gemma 4, plus planned premium model selection and LLM evaluation for improved reliability and analytics.
Hey everyone! ๐ Here's a quick update on what we've been working on over the past two weeks:
โ๏ธ CMX Insights & Inspiration
We recently attended CMX and had some great conversations with enterprise-level community managers. It was fascinating to see what large enterprises prioritize compared to smaller startups. These chats gave us some awesome inspiration for upcoming analytics features and a fresh new UI for the server listing!
๐ Revamped Landing Page
Weโve updated our landing pageโplease go check it out! They say a landing page is a company's manifesto, so we used it to showcase the features we've been perfecting, as well as our future focus areas: security & moderation: https://communityone.io/
๐ก๏ธ New Safety & Moderation CRM
We are actively building out a CRM system specifically for safety and moderation. This tool is designed for teams and offers extremely granular control. You will be able to:
- Assign custom strike values for different offenses.
- Use our analytics to proactively monitor chat or take automated actions.
- Easily manage strikes, timeouts, kicks, and bans.
Spark Update
We are resuming work on the Spark backend, but we already have a few exciting updates to share:
- New Web Widget: We built a web widget version of Spark that looks just like a Discord embed! The goal is to help you seamlessly drive more users from your website straight into your Discord server.
- Model Switch: Weโve updated our AI model, moving from OpenAI to Google (Gemma 4). It is a better model, matches the previous speed, and is going to be significantly more reliable. You can get more details on why we made the switch https://blog.communityone.io/switching-to-gemma-4-llm-evaluation/
๐ Premium Spark
For our Premium Spark users, we are preparing to roll out systematic LLM evaluation. We are also working on a new system that will allow you to select your preferred models, such as Gemini or Opus, whenever you want.
Let us know what you think, and as always, thank you for your feedback!
Best Discord AI Bot & Advanced Analytics Dashboard | CommunityOne
Fall back in love with Discord community building. AI-powered support, gamified quests, real-time analytics, and brand-grade moderation โ without the 14-hour grind.
Why We Switched to Gemma 4 โ And What Our Tests Actually Show
CommunityOne's Spark AI has run on OpenAI, then Gemini, and now Gemma 4 via OpenRouter. Here's the full technical breakdown: what drove each switch, our internal evaluation framework, and the actual test data behind the decision.
