Call /v1/ai/spam-score to review content risk and rewrite suggestions before sending. It consumes an AI credit on eligible plans; /v1/send does not automatically score or block messages based on this result.
Hourly analysis reviews eligible active workspaces against historical delivery metrics and may create deduplicated insights. Credit availability and plan eligibility apply.
Call /v1/ai/generate-template with an email brief to generate HTML, a subject, preview text and variables. Render the variables in your application, then send the resulting email.
Ask the assistant a natural-language question about your log feed — "why did opens drop yesterday afternoon" — and it returns a single-paragraph answer rooted in the actual logs, not a generic guess.
Request suppression suggestions to identify addresses worth reviewing. Your application decides whether to act; suggestions do not automatically suppress recipients.
A single short email summarising the week — what shipped, what regressed, what to fix this week. Specific to your data; no boilerplate.
Your AI integration checklist.
AI features — FAQ
Senddio uses its configured Anthropic integration for AI analysis and generation. Requested content and relevant log context are sent to the provider for that feature. Spam scoring is an explicit API call, not an automatic step in sending.
Free includes 20 AI credits per month, Starter has no included AI credits, Pro includes 1,000 credits, and Enterprise access is covered by your agreement.
On-demand features run when requested. Background features have separate eligibility and configuration; contact support before sending sensitive data if you need all AI processing disabled. This page does not promise per-feature dashboard opt-out controls.
AI requests send relevant content to the configured provider. Provider processing and retention are governed by its applicable terms and your agreement; do not assume zero retention or submit data you are not authorised to share.
Agent messages appear in Logs and relevant insight windows. Your application can request a spam score before sending; that analysis is separate from the send tool.