Upgrading to Erogen Premium (Silver and above) grants you access to special clusters of GPU servers dedicated to your membership tier. Higher tier memberships grant you access to higher resourced clusters, and will experience on average a shorter inference queue, shorter time-to-first-chunk, and faster tokens / second rate than lower tier clusters.

Note that due to normal traffic patterns, all clusters are subject to transient degradation.

Context Memory equates to the token limit that we use to measure the bot's recent memory. A token is basically a chunk of text that could be, for example, 1-4 characters long. So a memory limit of 2K would mean two thousand chunks of text can be included in the prompt that the AI sees when responding to your message, meaning that whatever chat history is included in that limit is visible to the AI and can be referenced in its new text generation.

You can think of Context Memory as short-term memory.

We try to replicate our memory model based on a simplified human memory model.

You know how memories from long ago are sparse and fuzzy? But there are a few that stick around and influence your current behavior more than most. Some people call those "core memories".

In our upcoming Core Memory Slots feature, we'll be providing a way for you to manually select bits of conversation and ask the AI to store that memory for reference for the long haul. We'll also be adding AI capabilities to the mix and allowing the AI to make its own decisions on what occurrences in your chat are memorable enough to store.

You can think of Core Memory Slots as long-term memory.

Our AI Speech-to-Text feature transforms your spoken language into written text messages to your AI chatbots in real-time. Perfect for users who prefer verbal communication or are multitasking, this tool captures every word with high accuracy. Whether you're dictating commands, sharing thoughts, or conducting a voice-driven session, our technology ensures that your spoken words are seamlessly converted into clear, actionable text.

Our AI Voice Triggers allow you to prompt your AI companions to speak aloud their messages to you. Character creators have the ability to select from different voices that match the personality of the character.

With AI Auto-Voice, your chatbot automatically converts written messages into natural-sounding spoken words without needing to be prompted. This feature allows users to experience interactions that feel more personal and engaging. Ideal for auditory people or those who enjoy podcasts and audiobooks, AI Auto-Voice brings your AI companion to life, speaking in a tone that fits the character's personality and context.

AI Phone Calls enable your chatbot to conduct real-time voice calls, making your AI companions as accessible as a phone call away. This feature integrates smoothly with your existing chats, providing users the convenience of talking to their AI directly without needing to type. Whether it's enjoying the festival season, frolicking under the spring blossoms, or just casual chatting, AI Phone Calls add a personal touch, enhancing user engagement through direct voice interaction.

Our Auto Avatar feature allows our character creators to add multiple images with descriptions to their character creations. This allows the AI to update their avatar live in conversation based on the recent chat history with you.

Auto Avatar initiates more frequently for higher tiers. For characters with many avatars - some can have up to fifty - this results in a highly dynamic roleplay experience.

Sometimes also referred to as "Spontaneous Messages", this feature allows our AIs to message you back without waiting for you to message them first.

Our current implementation of this feature is in beta and allows AIs to message Platinum users when it's been a certain amount of time since the last interaction. Typically, this will be an AI that you've interacted with very recently.

We plan on expanding this feature soon with double-texts and new introductions from AIs that you've never met before, as well as adding notifications so that you can better see when you've been messaged.

Our character creators take the act of crafting the future of AI companions really seriously. In order to measure performance and see what changes result in the most positive user impact, we've built analytics around character creator interactions that maintain the privacy of users that interact with your characters. We're going to be expanding our character creator analytics suite over time to create the most powerful character creator experience the world has ever seen.

User Personas are personality and appearance descriptions that you can don when in conversation with your AI companions. This will allow you to customize what your AI sees and interacts with in your chatbots without you having to explain everything in conversation. You can set these to be default personas per character.