Chat, Models & Context

Every message goes through Evelant's chat pipeline: your input is sanitized, checked against your plan's rate limit, optionally augmented with a live web search, and then sent to the model you've selected.

• Gemini is the primary routing path; ChatGPT and Claude model IDs are supported the same way once configured.
• The extension can attach scraped page text (and related metadata) to your request — you can turn page scanning off entirely in settings if you'd rather chat without page context.
• Image attachments are only sent for signed-in users, to keep guest usage lightweight and abuse-resistant.

See Multi-Model AI for the full model list and Plans & Limits for what each tier can access.