A regular conversation, fully equipped.
Responses stream in as they're generated, and the PFC you actually spend shows up after every reply. Templates, image attachments, web search, and voice input all live inside the same conversation.
What you can do from one input box
Get the extra work done without leaving the conversation.
Real-time streaming
Responses render token by token as the model generates them, so you're not staring at a spinner.
Auto model selection
Just type, and the app decides whether Chat, Compare, or Debate fits best — with a reason shown, and a manual override always available.
Input and output templates
Save frequent questions as templates with required variables, and lock output formatting with an output template.
Image attachments (Vision)
Drop in an image and the model reads it before answering — screenshots of documents work too.
Web search
Toggle on provider-native search from OpenAI, Anthropic, or Google to ground answers in current information.
Voice input
Speak into the microphone and it's transcribed straight into the composer — useful for long dictation.
PPTX generation
Ask for a deck and get a real, downloadable .pptx file. Word and Excel generation are also supported.
Automatic context management
Long conversations are trimmed and organized automatically so only the relevant context reaches the model.
Cross Review
Send the current answer to a different model for an independent review, then keep going with whichever one holds up.
How it works
Write
Type your question or instruction. Auto can decide which mode fits best.
Watch it stream
The answer streams in, and its actual PFC cost is recorded as it goes.
Add context
Combine templates, image attachments, voice input, and web search to sharpen the input.
Hand it off
Carry a promising answer into Compare or Debate, or request a Cross Review from another model.
Chat mode is ready to try the moment you create a free account.
Start free