Features
Switch models without switching apps.
Change providers mid-conversation, keep full context, and pick the model that actually fits what you're doing right now — reasoning, drafting, coding, or a quick answer.
- Full conversation history carries across model switches
- Per-conversation model memory — pick up right where you left off
- Model picker shows context window and rough cost per message
Rewrite the changelog entry so support can paste it as-is.
Working….
Fixed: exports larger than 50 MB no longer time out. Workspaces on the Pro plan can now export up to 2 GB.
Run one prompt across several models.
Send the same question to two or three models at once and read the answers side by side — useful the moment you're choosing a model for a new project, or just curious who's right.
Claude Opus 5
Ship the copy fix first — it's the cheapest change with the clearest signal from testing.
GPT-5.6 Sol
Fix the copy now, but flag the onboarding flow for a deeper redesign next sprint.
Built for real, ongoing work.
Projects
Group related conversations, files, and saved prompts together so context never gets lost between sessions.
File attachments
Drop in documents, images, or code and reference them across a conversation — even after switching models.
Prompt library
Save the prompts you reuse, organize them into folders, and fire them off with a keyboard shortcut.
Keyboard-first
A full shortcut set for switching models, jumping between conversations, and sending messages without touching the mouse.
Persistent history
Every conversation is saved locally and searchable — no expiring sessions, no lost context between launches.
Usage & cost view
See rough token usage and cost per provider so there are no surprises on your own API bill.