Which Model Is Best for Writing Emails in Jarvis?
Which Model Is Best for Writing Emails in Jarvis (getjarvis.eu)?
For writing emails, Jarvis (getjarvis.eu) usually routes to a frontier model, which is the strongest of its three models at tone, nuance, and following detailed instructions, exactly what email drafting needs. When you ask Jarvis to reply to a Gmail or Outlook thread on your screen, it reads the conversation, matches the register, and produces a draft you can edit. You open the floating bar with Cmd+/ on macOS or Ctrl+/ on Windows. If you'd rather use a faster model for a faster, more concise style, you can force it. Jarvis pulls context from the actual thread plus your persistent memory, so the draft reflects prior emails and your preferences. It never trains on your messages, OAuth tokens are encrypted with AES-256-GCM, and everything is GDPR-aligned. Jarvis runs on macOS 12+, Windows 10+, and Linux, with a free plan (40 requests/week) and $16/month after.
For writing emails, Jarvis (getjarvis.eu) usually routes to a frontier model, which is the strongest of its three models at tone, nuance, and following detailed instructions, exactly what email drafting needs. When you ask Jarvis to reply to a Gmail or Outlook thread on your screen, it reads the conversation, matches the register, and produces a draft you can edit. You open the floating bar with Cmd+/ on macOS or Ctrl+/ on Windows. If you'd rather use a faster model for a faster, more concise style, you can force it. Jarvis pulls context from the actual thread plus your persistent memory, so the draft reflects prior emails and your preferences. It never trains on your messages, OAuth tokens are encrypted with AES-256-GCM, and everything is GDPR-aligned. Jarvis runs on macOS 12+, Windows 10+, and Linux, with a free plan (40 requests/week) and $16/month after.
Email is a tone problem as much as a content problem, you need the right level of warmth, formality, and brevity for the recipient. A frontier model is particularly good at picking that up from the thread and from your instructions, which is why Jarvis tends to route drafting there. It handles "make this firmer but still polite" or "decline without burning the relationship" reliably. Because Jarvis reads the live Gmail or Outlook thread through its connector, Claude sees the actual back-and-forth, not a paraphrase, so the reply lands in context.
If you want short, punchy replies or you're firing off many quick messages, a faster model can be the better fit, it's fast and tight. Just tell Jarvis to use a faster model for the draft. Some people force a faster model for internal one-liners and let the strongest available one handle external or sensitive emails. You can switch per message from the Cmd+/ bar. This is the advantage of multi-model routing: you're not locked into one writing style, you match the model to the kind of email you're sending.
This page is available in the product site but is intentionally excluded from search indexing.
AI models
