AI drafts specific violation notices with evidence and appeal instructions. Your trust & safety team reviews every notice before delivery.
Vague moderation notices frustrate users and generate support tickets. When users only see "your content was removed" without specifics, they appeal everything — creating more work. Writing specific, fair notices for each violation type is time-consuming, and inconsistent language across moderators creates legal risk.
MultiMail's AI agent drafts moderation notices that cite the specific policy section violated, describe the content that triggered the action, and include clear appeal instructions. With gated_send oversight, your trust & safety team reviews every notice for accuracy and fairness before delivery.
When your platform flags or removes content, the event triggers MultiMail's AI agent. The agent receives the violation type, evidence, and user details.
The AI composes a notice citing the exact policy section, describing what was flagged, and explaining the action taken. It includes appeal instructions and timeline based on the violation severity.
With gated_send, every moderation notice is reviewed by your trust & safety team before delivery. They verify the violation citation is accurate and the tone is professional.
Approved notices are sent. The agent monitors replies for appeal requests and routes them to the appropriate review queue using tag_email.
import requests
API = "https://api.multimail.dev/v1"
HEADERS = {"Authorization": "Bearer mm_live_xxx"}
def send_moderation_notice(user_email: str, violation: dict):
body = (
f"Your content titled &"cm">#039;{violation['content_title']}' was "
f"{violation[&"cm">#039;action']} for violating our "
f"{violation[&"cm">#039;policy_section']} policy.\n\n"
f"Specifically: {violation[&"cm">#039;details']}\n\n"
f"To appeal this decision, reply to this email within "
f"{violation[&"cm">#039;appeal_window_days']} days with additional context."
)
response = requests.post(
f"{API}/send",
headers=HEADERS,
json={
"from": "[email protected]",
"to": user_email,
"subject": f"Content policy notice: Your post was {violation[&"cm">#039;action']}",
"text_body": body
}
)
return response.json()
send_moderation_notice("[email protected]", {
"content_title": "Special offer",
"action": "removed",
"policy_section": "Spam (Section 3.2)",
"details": "Post contained promotional links without disclosure",
"appeal_window_days": 14
})Generate a policy violation notice with specific details and appeal instructions.
import requests
API = "https://api.multimail.dev/v1"
HEADERS = {"Authorization": "Bearer mm_live_xxx"}
"cm"># Check for appeal replies
inbox = requests.get(
f"{API}/mailboxes/trust-safety-mailbox-id/inbox",
headers=HEADERS,
params={"status": "unread"}
).json()
for message in inbox["emails"]:
email = requests.get(
f"{API}/emails/{message[&"cm">#039;id']}",
headers=HEADERS
).json()
# Check if this is a reply to a moderation notice
thread = requests.get(
f"{API}/threads/{email[&"cm">#039;thread_id']}",
headers=HEADERS
).json()
if is_moderation_thread(thread):
# Tag as appeal for review queue
requests.post(
f"{API}/emails/{email[&"cm">#039;id']}/tags",
headers=HEADERS,
json={"tags": ["appeal", "pending-review"]}
)Monitor for appeal replies and route them to the appropriate review queue.
"cm">// Content moderation notice via MCP tools
"cm">// 1. Send moderation notice
await mcp.send_email({
from: "[email protected]",
to: userEmail,
subject: "Content policy notice: Your post was removed",
text_body: `Your post titled '${contentTitle}' was removed for violating our Spam policy (Section 3.2). To appeal, reply within 14 days.`
});
"cm">// 2. Check for appeal replies
const inbox = await mcp.check_inbox({
mailbox_id: "trust-safety-mailbox-id",
status: "unread"
});
"cm">// 3. Tag appeals for review
for (const msg of inbox.emails) {
const thread = await mcp.get_thread({ thread_id: msg.thread_id });
if (isModerationThread(thread)) {
await mcp.tag_email({
email_id: msg.id,
tags: ["appeal", "pending-review"]
});
}
}Use MultiMail MCP tools for content moderation notice workflows.
Specific, well-written notices that explain exactly what was violated and why reduce frivolous appeals by giving users clear information upfront.
AI drafts from approved policy templates ensuring every notice uses consistent language, reducing legal risk from inconsistent moderation communication.
Every notice includes specific evidence, the exact policy section, and clear appeal instructions — meeting regulatory requirements like the Digital Services Act.
Trust & safety reviews every moderation notice before delivery, catching edge cases where the AI might apply the wrong policy or mischaracterize the violation.
Email infrastructure built for AI agents. Verifiable identity, graduated oversight, and a 38-tool MCP server. Formally verified in Lean 4.