Inbound email API · webhooks, REST, MCP
An email inbox your AI agent can run.
Inboundr is real email on your own domain: parsed and threaded, delivered by signed webhook, read over REST or MCP.
Free forever tier · No credit card required
The whole integration is one command.
There's no mail server to run and no MIME to parse.
Try it now
A real inbox, right here.
No signup. We'll mint you an address on our domain — email it, and watch it arrive parsed, verified, and shaped exactly as your agent would receive it.
Create a real inbox on our domain, then email it.
It runs on the same pipeline your production mail would.
Demo inboxes expire after 15 minutes.
Use cases
Receive email on every address, instantly.
Point a domain at Inboundr and it starts accepting mail on any address at that domain — none of them need creating first.
Capabilities
Everything an inbound email API has to do.
Receiving is the easy half. These are the parts that usually take a month of your own code.
01 · Webhooks
Parsed email arrives as a signed webhook
Every message is parsed and POSTed to your app, HMAC-signed so you can prove it came from us.
02 · REST API
Read the inbox over the REST API
Unrouted addresses just store their mail, ready to fetch over REST or read by your agent over MCP.
03 · Shield
Filter mail before it lands
Block, allow, route or quarantine by sender, subject or domain — rules run before anything reaches your webhook.
04 · Delivery
Deliveries that keep trying
Failed webhooks retry on a backoff you set, and every attempt is logged with its response for replay.
05 · Routing
Claim a whole domain
One catch-all routes every address you never explicitly created, so agents can provision their own.
06 · Attachments
Attachments without extra storage
Every attachment gets a short-lived download link over REST or MCP — no bucket of your own to wire up.
07 · DMARC
Turn rua= reports into a trend chart
Point rua= and tlsrpt= at us, and every report becomes pass-rate history — plus an alert the day a new source starts failing.
Model Context Protocol
Eleven MCP tools your agent can call.
Every capability is a callable tool — search them the way you'd search any other command palette.
Domains & inboxes
- list_domainsSee your domains and whether they're verified.
- create_inboxMint a new address on your domain, live instantly.
- list_inboxesEvery address the agent owns, at a glance.
Reading
- list_emailsScan the inbox newest-first, filtered and paginated.
- get_emailOne message in full: bodies, attachments, verdicts.
- get_threadThe whole conversation, oldest first.
- get_attachmentA short-lived download link for any attachment.
Shield
- create_shield_ruleBlock, allow, route or quarantine inbound mail.
- list_shield_rulesThe filter rules, in the order they fire.
- list_shield_blocksWhat got dropped, and which rule caught it.
- delete_shield_ruleRetire a rule; its senders flow again.
How it works
Set up inbound email in three steps.
From a domain you own to an agent running its own inbox.
- 01
Point your domain
Add a domain, drop in the MX and DKIM records we generate, and we verify it for you. Every address on it is now yours.
- 02
Connect your agent
One command adds the MCP server and every tool above is live. Prefer HTTP? The same endpoints are a plain REST API with a Bearer key.
- 03
Let it run the inbox
Mail arrives parsed and HMAC-signed. Your agent reads the thread, provisions more addresses, and replies in-thread.
Pricing
Pay for the mail, not the seat.
One monthly allotment for what arrives, and nothing charged per seat, per inbox, or per agent.
Free
Kick the tyres.
- 200 emails / mo
- Unlimited inboxes
- 1 domain
- 50 emails / day
- No attachment storage
- 3 Shield rules
- No DMARC/TLS-RPT monitoring
- Stops at the allotment
Basic
One agent, in production.
then $0.60 per 1,000
Get started- 6,000 emails / mo
- Unlimited inboxes
- 3 domains
- No daily limits
- 10 MB attachments
- 10 Shield rules
- DMARC/TLS-RPT monitoring
- Pay-as-you-go past the allotment
Pro
A fleet of agents.
then $0.60 per 1,000
Get started- 50,000 emails / mo
- Unlimited inboxes
- 50 domains
- No daily limits
- 25 MB attachments
- 50 Shield rules
- DMARC/TLS-RPT monitoring
- Pay-as-you-go past the allotment
Scale
High volume.
then $0.60 per 1,000
Get started- 150,000 emails / mo
- Unlimited inboxes
- 500 domains
- No daily limits
- 50 MB attachments
- Unlimited Shield rules
- DMARC/TLS-RPT monitoring
- Pay-as-you-go past the allotment
Receiving gets its own monthly allotment, and stops at 100%. Turning on pay-as-you-go keeps mail flowing, up to a monthly limit you set.
- Off by default
- Nothing above your plan price until you enable it
- Capped by default
- And the cap is the gap to the next plan, so it can never cost more than upgrading
- Warned first
- We email at 80% and 90%, before anything stops
- Every plan
- MCP, REST API, webhooks, unlimited inboxes, 15-day history
Monthly billing, cancel anytime.
How much do I have to set up?
Just a domain you control. Add it, paste in the MX and DKIM/SPF records we generate, and the console verifies them automatically; you're receiving your first parsed email in about five minutes. No mail server, no parser, nothing of yours left running.
Do I have to use MCP?
No. MCP is the shortcut, not a requirement. Everything it exposes is also a plain REST API with a Bearer key, and inbound mail arrives as a signed webhook either way. Start with whatever your stack speaks today and switch later without rebuilding anything.
Can agents create their own addresses?
Yes, that's the point. An agent calls create_inbox on any verified domain and is receiving mail at that address immediately. Inboxes are unlimited on every plan, including Free. Make one per customer, per task, or per agent without asking us first.
Does it send email too?
No — Inboundr is built for the receiving half: parsing, threading, and routing mail to your webhook, REST API, or MCP agent. Pair it with whatever you already send through; there's nothing to migrate on that side.
What does an inbound email look like to my code?
Clean JSON, never raw MIME. Text and HTML bodies, attachment metadata, threading headers, and the SPF/DKIM/spam verdicts, over MCP, REST, or a webhook that's HMAC-signed so you can prove it came from us. Your agent reads email the way it reads any other API.
What happens when I hit my plan's limit?
Nothing you didn't agree to up front. We email you at 80% and 90%, and at 100% receiving pauses until the cycle resets. Paid plans can switch on pay-as-you-go to keep going instead, billed per 1,000 emails on that month's invoice. It's off until you turn it on, and when you do it comes with a monthly spending limit already set. By default that's the difference to the next plan up, so pay-as-you-go can never cost you more than upgrading would. Change it, or remove it, whenever you like.
When does my allotment reset?
On your own anniversary, not the 1st of the month. Start on 12 July and your cycle runs 12 Jul → 12 Aug → 12 Sep, matching exactly what you're billed for. Your allotment resets on that date, and nothing carries over.
How does this compare to the alternatives?
There is a page for each of them, and each one ends with the case for choosing them instead: sending platforms with inbound attached, forwarders, and building it yourself. For some workloads those genuinely are the right answer — the comparisons say so where they are.
How hard is it to leave?
One click, no lock-in. Billing is monthly through Stripe's portal, with no annual contract. The domain and addresses are yours, so nothing about your email identity is stuck with us. Cancel and you drop to Free: your domain keeps receiving, within the Free allotment.
Free tier open
Stop building the email layer.
Point a domain, add the MCP server, and your agent is running its own inbox in about five minutes. First domain is free.
Free forever tier · No credit card required