AsheNow brings conversations, models, and configurable Agents into a desktop workspace. These are capabilities in the application under development, not a promise of public availability.
Start and revisit conversations, watch streaming responses, and work with Markdown, code, and rich text. Conversation panes can sit side by side or stack so related work stays in view.
Create Agents with a guided setup, then fine-tune them in an editor organized into Profile, Brain, Instructions, Abilities, and Advanced. Give each one an avatar you can change any time. Add Agents to a conversation and choose whether each responds after you, after another Agent, manually, or silently.
Choosing a model is a walk-through, not a guess: first where the Agent runs, then which model, with recommendations and the real price per million tokens. If a model does not support a setting such as temperature or reasoning effort, AsheNow adjusts it automatically so replies keep working.
Tell an Agent what it should make: chat and text, images, or video. Image Agents paint pictures from a description, and can restyle a photo you attach with models that accept pictures. Video Agents render short clips; they take a few minutes and keep going if you close the chat. Shape, resolution, length, and cost are shown before you start.
Press Listen on any reply to hear it, or click the microphone to dictate a message. Use OpenRouter voices, or free voices that run on your own PC, offline, so the text never leaves the computer. Each Agent can have a voice of its own.
Connect OpenRouter, DeepSeek Direct, Anthropic, and custom endpoints that implement supported OpenAI Chat Completions or Anthropic Messages protocols. API keys are encrypted with Windows protection and can be shared across Profiles or kept to one.
Organize conversations and Agents into Profiles, with provider connections and preferences scoped for the way you work. Profiles, conversations, and generated files are stored in the local workspace.
Group related conversations and Agents in Projects. Attach pictures, audio, and documents for models that can read them; images, audio, and video show right in the conversation. Attachments stay connected to their conversation or Project.
A game mod can let its characters talk through AsheNow. Each character becomes an Agent with its own conversation, so it remembers you, and replies stream back one sentence at a time with speech. The bridge only answers on your own PC and needs a key you create in Settings. The first mod, for Skyrim, is in development.
Image, video, and OpenRouter voice features use OpenRouter and are billed by it per use. Provider support depends on the documented protocol and capabilities of a connection. A custom endpoint URL does not imply that every model, API, file type, or provider feature is supported.
Public downloads are not available yet. Join the list and tick the beta box to be considered when testing begins. Invitations are not guaranteed.
