Protect deliverability at the platform level

MultiMail monitors bounce rates, complaint rates, and domain health so your agent sends from clean infrastructure — without manual intervention.


Why this matters

Email deliverability breaks quietly. New mailboxes get throttled by ISPs before they've built reputation. Domains on shared IP ranges inherit bad neighbors' spam scores. A single campaign with a 0.5% complaint rate can trigger a block that takes weeks to reverse. Most AI agents have no visibility into these signals — they keep sending until the damage is done, at which point the remediation window has already closed.


How MultiMail solves this

MultiMail enforces deliverability guardrails at the infrastructure layer. New mailboxes have gradually increasing daily send limits that prevent fresh senders from hitting volume thresholds before ISPs trust them; exceeding the current limit returns HTTP 429 rather than silently dropping the send. Custom domains are verified for SPF, DKIM, MX, and return-path alignment via GET /v1/domains/{id} before they can send. Every outbound message carries a signed X-MultiMail-Reputation header, and any recipient can look up that reputation at the public GET /.well-known/reputation/{hash} endpoint. Sustained bounce and complaint problems lower a mailbox's standing and tighten its limits automatically — protecting the shared sending commons without requiring operator intervention in the critical path.

1

Warmup within safe limits

MultiMail enforces per-mailbox daily send limits during the warmup period. New mailboxes ramp from 50 to 500 to 5,000 emails per day over 30 days, matching the volume curve ISPs expect from legitimate senders. Exceeding the limit returns HTTP 429 with a Retry-After header — the send is not silently dropped.

2

Verify the sending domain before sending

Before sending from a custom domain, your agent queries GET /v1/domains/{id} to confirm SPF, DKIM, MX, and return-path alignment. The response includes the per-record verification flags (dkim_verified, spf_verified, mx_verified, return_path_verified) and the dns_records you must publish. A domain that has not reached verified status cannot send.

3

Send with verified identity

The mailbox send endpoint (POST /v1/mailboxes/{mailbox_id}/send) sends from a mailbox on a verified domain. MultiMail signs outbound messages with DKIM, attaches the signed X-MultiMail-Identity claim, and enforces alignment on the sending domain. The sender is the mailbox in the path plus your bearer key — there is no from field to spoof, so a different domain cannot be impersonated.

4

Carry a verifiable reputation

Every outbound message carries a signed X-MultiMail-Reputation header — a hash that any recipient can resolve at the public GET /.well-known/reputation/{hash} endpoint to see the sender's standing without trusting the From header. Reputation is computed from real delivery outcomes (bounces, complaints, and spam-trap hits), so building a clean send history directly raises how recipients and ISPs treat your mail.

5

Automatic enforcement on sustained problems

Google and Yahoo's 2024 bulk-sender requirements cap complaint rate at 0.1% and bounce rate at 2%. As an account's bounces and complaints accumulate, MultiMail lowers its reputation score and raises its enforcement tier — which tightens send limits and gates higher-trust actions like provisioning additional mailboxes. Sustained abuse degrades standing automatically, no operator action required.


Try it with your agent

Pick your platform, copy the prompt, and paste it to your AI agent — it sets up MultiMail and builds the whole flow. Nothing to fill in.

1. Read https://multimail.dev/llms.txt, connect the MultiMail MCP server, create a free inbox, and set up a verified sender. 2. In HubSpot, use Workflows enrollment triggers for contacts entering active marketing segments, lifecycle stages, form submissions, or list membership changes that normally start outbound email. 3. Before each send sequence, inspect contact engagement, suppression status, recent hard bounces, unsubscribes, and campaign volume; split new or risky audiences into small daily batches, personalize by lifecycle stage and last known interest, and pause any segment showing elevated bounce or complaint risk. 4. Send or schedule through MultiMail in autonomous oversight mode, respecting gradual mailbox ramp-up and letting MultiMail tighten limits when reputation signals deteriorate. 5. Keep everything in draft/test mode until you ask me only for HubSpot access, the sending brand voice, and permission to go live.

What you get

Warmup enforcement before sends reach ISPs

MultiMail applies daily send limits at the API layer during warmup. New mailboxes ramp volume automatically over 30 days without manual scheduling. Attempts to exceed the limit return a 429 — the send is never silently dropped or queued past the safe window.

Automatic enforcement on sustained problems

Google and Yahoo enforce a 2% bounce and 0.1% complaint ceiling for all bulk senders. As an account's bounces and complaints accumulate, MultiMail lowers its reputation score and raises its enforcement tier — tightening send limits and gating higher-trust actions automatically. Your agent does not need to poll for these conditions; enforcement is a platform-level guarantee.

Verifiable sender reputation

Every outbound message carries a signed X-MultiMail-Reputation header derived from real bounce and complaint outcomes. Any recipient can resolve it at the public GET /.well-known/reputation/{hash} endpoint without trusting the From header — so a clean send history is reputation any party can verify, not a private score.

Domain verification before sending

SPF, DKIM, MX, and return-path alignment is verified per domain via GET /v1/domains/{id} and POST /v1/domains/{id}/verify. A domain that has not reached verified status cannot send, so misconfigured domains never reach ISPs and damage sender reputation across the shared pool.

CAN-SPAM compliance enforced at the infrastructure layer

MultiMail enforces unsubscribe processing, opt-out honoring, and suppression list management as infrastructure constraints under CAN-SPAM. Your agent does not need to implement these requirements in application logic — non-compliant sends are rejected before delivery.


Recommended oversight mode

Recommended
autonomous
Deliverability optimization is reactive and time-sensitive. ISP signals arrive continuously and require immediate adjustment — warmup throttles, reputation enforcement, and domain health blocks all need to respond faster than a human approval loop allows. The enforcement logic is deterministic and platform-enforced by MultiMail, so the risk surface of autonomous operation is bounded by platform guardrails rather than agent judgment. Human oversight belongs at the policy level — setting custom thresholds, reviewing suspended mailboxes, and approving domain configuration changes — not in the per-send loop where latency directly degrades deliverability outcomes.

Common questions

What happens when an account's complaint rate climbs?
As complaints accumulate past the Google and Yahoo enforcement threshold (0.1%), MultiMail lowers the account's reputation score and raises its enforcement tier, which tightens send limits automatically. Sustained abuse keeps degrading standing until limits are very restrictive. Recovery comes from improving real delivery outcomes — pruning bad addresses, fixing content, and letting a clean send history rebuild the score over time.
How does warmup work for new mailboxes?
New mailboxes are subject to a daily send limit that increases over a 30-day warmup window: 50 per day in week 1, 500 per day in week 2, 2,500 per day in week 3, and up to 5,000 per day in week 4. Attempts to exceed the limit return HTTP 429 with a Retry-After header indicating when capacity resets. The ramp schedule matches what major ISPs expect from a new sender building legitimate reputation.
Does MultiMail handle DKIM signing automatically?
Yes. When you add a custom domain via POST /v1/domains, MultiMail provides the DNS records to publish (returned in dns_records). All outbound messages from that domain are signed automatically once verified. You can confirm alignment using GET /v1/domains/{id} before starting a campaign — the response includes dkim_verified, spf_verified, mx_verified, and return_path_verified flags and the overall status.
How does sender reputation get exposed?
Reputation is not a private score you poll — it travels with your mail. Every outbound message carries a signed X-MultiMail-Reputation header containing a hash, and any recipient can resolve that hash at the public GET /.well-known/reputation/{hash} endpoint (rate-limited to 10 lookups/hour) to see the sender's standing without trusting the From header. The standing is derived from real bounce and complaint outcomes, so a clean send history is reputation any party can independently verify.
Can I enforce a stricter deliverability bar than the platform floor?
Yes — in your own agent. The platform enforces the Google and Yahoo minimums (0.1% complaint rate, 2% bounce rate) as a non-negotiable floor via reputation scoring and enforcement tiers; you cannot loosen those. To hold a tighter bar, track your own bounce and complaint outcomes from email.bounced webhooks and pause your send queue earlier — for example at 0.05% complaint rate — before MultiMail's platform floor would ever trigger.
How does MultiMail distinguish hard bounces from soft bounces?
Hard bounces (permanent delivery failures indicated by 5xx SMTP codes) are immediately added to your tenant's suppression list. That address is refused with a 400 on the next send, and hard-bounce suppressions are permanent (DELETE /v1/suppression/{address} returns 403 for them). Soft bounces (temporary failures, 4xx codes) are retried with exponential backoff. Only permanent failures count toward the bounce signal used for reputation enforcement.

Explore more use cases

The only agent email with a verifiable sender

Email infrastructure built for AI agents. Verifiable identity, graduated oversight, and a hosted MCP server. Formally verified in Lean 4.