Email is sensitive on both sides: reading exposes private content, sending speaks on your behalf. AgentRow lets you mark the read tool sensitive, put the send tool behind approval, and block topics in the final answer.
An assistant that summarises a thread the user pastes or that a read tool fetches, and drafts a reply in the house tone.
A support agent that sends a confirmation email through a transactional provider after the user approves the text.
An internal agent that turns a chat outcome into an email digest for a team alias.
The shape a tool takes in the AgentRow builder API. Credentials are stored on the tool's auth settings, not in its configuration, so they never reach a prompt. Placeholders in double braces are filled by the model from the input schema.
{
"name": "send_email",
"type": "http_api",
"description": "Send an email. Use only after the user approved the recipient, subject and body.",
"auth_type": "api_key",
"safety_level": "write_sensitive",
"requires_approval": true,
"config": {
"method": "POST",
"url": "https://api.example-mail.com/v1/messages",
"headers": { "Content-Type": "application/json" },
"body_template": { "to": "{{to}}", "subject": "{{subject}}", "text": "{{text}}" },
"allowed_domains": ["api.example-mail.com"]
},
"input_schema": {
"type": "object",
"properties": {
"to": { "type": "string", "description": "Recipient email address" },
"subject": { "type": "string" },
"text": { "type": "string", "description": "Plain-text body" }
},
"required": ["to", "subject", "text"]
}
}Every call appears in the run trace with the rendered request, the response status and body, timing, retries and errors, plus what the model was shown of the result. If the model never called the tool, the toolbox step shows whether it was bound for that model call and why.
The run pauses before the call executes and the chat shows an approval card with the tool, its arguments and its safety level. Approve runs the call; reject resumes the run with an explicit refusal the model sees. The decision is recorded on the tool call.
Blocked topics and deny or allow rules are checked on the final answer, and apps add their own safety check after the agent's. Put the rules in the agent's Guardrails tab and the app's settings.
An email tool type exists but is not executable yet; HTTP API tools against the Gmail API or a transactional provider are the supported path today.
Connect AI agents to HTTP APIs, webhooks, web pages and other agents with typed tools: JSON schemas, auth types, safety levels, limits and call logs.
Tool types, what a tool defines, how results reach the model, tool selection with search_tools, and authenticated tools for signed-in users.
AgentRow vs Relevance AI for AI agents: workflow control, multi-agent apps, tools, knowledge, observability and production readiness, compared.
Slack, HubSpot, Stripe, Shopify, GitHub, calendar, email, wikis, webhooks and any REST API.
Start a free trial and wire it up, or tell us which systems you need to connect and we will point you at the right tool definitions.