Skip to main content

​
Choose a model

Open your Verb’s AI Engine. The model selector shows the models available to your plan and their credit costs. Start with an included free option; switch when you need a different style or capability. The text catalog refreshes daily with current models, prices, and provider logos. Higher plans include lower-plan models; check the live selector for current access and costs. Paid-model usage still consumes credits. Free includes Valley 3 and the active Model of the Week, which costs no model credits during its free week. See Plans and credits for allowances and how the weekly selection works.

​
Tune its replies

You can leave the defaults alone. Change a setting only when you know what you want to improve: Higher context and longer replies can use more credits on paid models. Recent context is different from long-term Memory. There is no plan-based output-token cap. Your reply-style settings, the selected model/provider’s limits, and service timeouts still determine how long a reply can be.

​
Mood and social state

Social state controls qualities such as energy, focus, playfulness, and irritation. You can set a starting state and allow it to evolve with conversations. These controls affect tone; they aren’t a record of a person’s real emotions.

​
Autonomous replies

Autonomous Mode lets your Verb decide when to join eligible conversations without a direct request. Use a lower frequency in busy channels. Platform permissions and channel activation still apply—turning it on doesn’t grant access to every channel. Enable Web Search when your Verb needs current information. It searches when relevant, rather than on every message. Ask for sources when accuracy matters and check them yourself.

​
Response streaming

On Ultra, Response Streaming can reveal replies as they are generated in supported integrations, including Discord, WhatsApp, and Telegram. It doesn’t change the character’s personality. API streaming is a separate request option; the dashboard toggle doesn’t enable it for API calls.