Models
Wayfind runs on Meta's AI models — one for chatting, one for images, one for voice. You never pick models or paste API keys; Wayfind routes each job to the right one and the spend lands in your usage.
| Job | Model | What it does |
|---|---|---|
| Chat | Muse Spark Open | Every conversation, tool call, and routine on all plans. |
| Fresh facts | Muse Spark + web grounding | Bots search the web and answer with cited sources. |
| Images | Muse Image (muse-image-1.0) | Generate and edit images from a text instruction. |
| Voice | Muse Voice Transcribe | The mic button in chat turns speech into the job. |
Chat
Muse Spark Open serves every conversation on all plans; Premium adds Muse Wonder for browser work and heavier agentic tasks. You describe the outcome — the routing is Wayfind's job. See Billing for how chat spend works.
Image generation
Ask any Bot to generate or edit an image — “a hero banner for the spring sale” or, of an existing image, “make it night”. Muse Image renders it, the picture lands inline in chat and in the workspace generated/ folder. Flat $0.01 per image, billed only on success.
Voice
Tap the mic button in chat and speak the job instead of typing it. Muse Voice transcribes the recording and the Bot carries on as if you'd written it.
Grounded search
When a Bot needs facts fresher than training data — prices, docs, today's news — it searches the web through Muse Spark and answers with a Sources list, so you can verify anything that matters.
Learn more
- Muse Spark — chat and reasoning.
- Muse Image — generation and editing.
- Muse Voice Transcribe — speech to text.