RadMail is an AI email operating system that reads every inbound message, surfaces only what matters right now, and drafts the follow-ups you promised on the day they are due.
> the inbox that thinks — and follows through.
[ the BEC hard-stop ] Moving money, changing banking, and first contact are human-only — forever. A built-in defense against business-email compromise no other inbox will make.
The three emails that matter, surfaced first. RadMail ranks every message on importance andurgency, drafts the follow-ups you promised on the day they're due, and tracks what actually earned a reply — not just an open.
built for regulated industries
Where a misrouted reply or a spoofed wire request isn't an oops — it's a compliance event. RadMail makes moving money, changing banking, and first contact human-only, forever.
Regulated industry? Talk to us ▸
It does the inbox work, not just speed it up.
Importance is not urgency.
Two separate axes. A loud newsletter scores low; a quiet COA sign-off scores high. Every message gets a plain-English reason it landed where it did.
A Right Now lane you can trust.
One short list of what actually needs you, this minute — the rest stays out of your way. Per-sender learning sharpens it over time.
It remembers what you promised.
RadMail tracks every commitment in your threads — owed by you and owed to you — drafts the reply the day it's due, and watches for the response so it never nags about what's already done.
It knows what actually earned a reply.
Because RadMail reads inbound mail as well as sending it, it ties every send to the reply it produced — not just an open, which a send-only tool can't see past.
Watch a thread. It knows who has the last word.
Mark a thread watched: when the other side spoke last it pins to the top of Right Now until you answer; when you spoke last it waits in a quiet Watching panel — visible, never nagging, and never tunable into silence.
Money, changed banking, and first contact with a stranger are human-only. Forever. RadMail will never hand you an auto-sendable reply for those — a deliberate defense against business-email-compromise fraud.
Enforcement model: capability-absent. Not a promise we keep — a tool we never built. The three ways an agent product can restrain itself, and which one this is:
capability-absent← RadMail — The tool that would perform the forbidden action does not exist on the server, so it cannot be called. There is nothing to configure and nothing to bypass.config-restricted— The capability exists and is narrowed by configuration — an allowlist, a scope, a policy file. Whoever holds the configuration can widen it again.policy-gated— The capability exists and is restrained by instructions, prompts, or documented policy that a model is asked to follow.
No setting, feature flag, plan tier, environment variable, or admin toggle on RadMail's MCP server enables a forbidden action. Configuration can widen what an agent READS — RADMAIL_API_KEY and RADMAIL_API_URL unlock connected mode, which is read-only. Disclosed here rather than left for you to find: the package's public MIT source gained an opt-in, default-OFF send tool on 2026-10-03 (RADMAIL_SEND_TOOL=1 plus a SEPARATE send-scoped RADMAIL_SEND_API_KEY), which is not in the published 0.5.0 build. It does not widen the forbidden set. It hands a message to RadMail's outbound gate, where the hard-stop classes are held for a person, and it cannot release a hold — its client has exactly one endpoint and the API has no release endpoint for it to call. The hosted sandbox and every hosted HTTP entry construct the server with no send option at all, so on that tier no environment variable can add one. Enabling a forbidden action would still require shipping a tool that does not exist.
Don't take our word for it — the contract is machine-readable at /.well-known/agent-safety.json, and the sandbox tool list is public, so you can check that no such tool exists.
For inboxes that can't afford a mistake.
In a regulated business a misrouted reply or a spoofed payment request isn't an oops — it's a reportable event. The triage engine and the BEC hard-stop — no tool can move money, change banking, or make first contact — work in every inbox today. The per-industry compliance pack is what plugs in, and new regimes ship as PolicyPacks. Same engine from a solo practice to a whole firm.
○ compliance pack on the roadmap — the BEC hard-stop & triage are live for all today
PHI lands in the inbox — a reply to the wrong address is a breach.
- The message that needs a clinician now, surfaced first.
- Patient and referral follow-ups drafted the day they're due.
Wire requests and account changes are the BEC bullseye.
- Money and banking-change emails stop cold — human-only.
- The client message that can't wait, surfaced first.
CUI handling on BAA-covered, audited infrastructure.
- Solicitation and RFP deadlines tracked — nothing slips.
- First-contact and payment changes stay human-only.
Trust-account wire fraud is the bar's recurring nightmare.
- A changed-wire instruction can never auto-send — human-only.
- Client follow-ups and filing deadlines drafted on time.
Closing wire fraud is a multi-billion-dollar BEC problem.
- A spoofed 'new wiring instructions' email never auto-sends.
- Buyer, lender, and closing follow-ups drafted the day they're due.
Client financials and refund-routing requests are a fraud target.
- Payment-detail and bank-change emails stay human-only.
- Deadline and client follow-ups drafted the day they're due.
Claims and policyholder PII move through the inbox all day.
- The claim or broker message that can't wait, surfaced first.
- Payout and banking-change requests stop cold — human-only.
More on the way:pharma & clinical research, education (FERPA), and any inbox where a misdirected reply or a spoofed payment request is a regulated event. Compliance is a shared responsibility — RadMail is a tool, not a guarantee, and is not HIPAA-certified or FedRAMP-authorized.
Press play. Watch the buried urgent wire jump to the top.
A faithful simulation of the live triage engine reading a raw inbox. Scripted re-enactment, not a live model call — the engine itself is live in a test bed on two real businesses.
Compliance is a pluggable per-organization regime assembled from PolicyPacks. We sign a BAA and process your data on BAA-covered, encrypted, audited infrastructure. Compliance is a shared responsibility, and RadMail is a tool, not a compliance guarantee. It is not HIPAA-certified or FedRAMP-authorized. The SOC 2, HIPAA, and FedRAMP/CUI packs are architecture and roadmap.
Launch pricing. Try it free now— paid plans turn on at launch. Cheap to start because it's cheap to run; priced for real value where the stakes are real.
Regulated
For inboxes that carry regulated data and can't afford a mistake.
- Everything in Business
- Compliance PolicyPacks (HIPAA, finance, legal)
- Signed BAA · BAA-covered, encrypted, audited infra
- Dedicated database + reads-audit log
- Permanent BEC hard-stop on money / banking / first-contact
Business
most popularEmail that handles itself — drafts everything, sends the safe ones.
- Everything in Personal
- Autonomous safe-send (provably-safe replies only)
- Multiple inboxes + team
- SOC 2-baseline regime
- Priority support
Personal
Your inbox, triaged by what actually matters.
- 1 inbox (Gmail or Outlook)
- Two-axis triage + the Right Now lane
- Commitment tracking — never drop a reply
- AI reply drafts, ready for you to review
Free / Agent
Point an agent at the sandbox, or try the live triage demo.
- Heuristic triage engine, in memory
- The MCP surface for AI agents
- Explainable 'why surfaced'
- No signup required
Enterprise — SSO, audit export and custom terms. Contact sales.
Compliance is a shared responsibility and RadMail is a tool, not a compliance guarantee — not HIPAA-certified or FedRAMP-authorized. The SOC 2, HIPAA, and FedRAMP/CUI packs are architecture and roadmap.
Open to founding teams.
The engine is live today in a test bed on two real regulated retailers. Run the engine, then join the waitlist — shape what ships and lock founding terms before launch.
Questions, answered straight.
What is RadMail, and how is it different from Superhuman or an AI inbox?
- Those are clients you read mail in — a faster inbox. RadMail is an email operating system that does the work: it ranks each message by importance and urgency on two separate axes, surfaces a Right Nowlane with a plain-English reason it landed there, extracts the commitments hiding in your threads, and drafts the follow-up the day it's due. And it tells you what actually got a reply, not just an open — a send-only tool can't, because it never sees the replies. RadMail is a Sureel venture.
What is reply-correlation?
- Most email tools report opens and clicks. RadMail can tell you which of your sends actually earned a reply, because it ingests inbound mail as well as sending outbound — a send-only email service provider never sees the reply, so it can't connect a send to the response it produced. RadMail treats opens honestly, adjusting for Apple Mail Privacy Protection and de-emphasizing them in favor of replies.
Does RadMail send emails on its own?
- Only after it earns trust, and never for high-risk actions. RadMail rolls out in stages: surface-only, then draft-on-due (it writes the message and waits for you), then narrow auto-send for low-risk cases. Anything involving money, banking details, a decision, or first contact with a third party stays human-only, forever — a deliberate defense against business-email-compromise fraud.
Is RadMail HIPAA or SOC 2 compliant?
- Compliance is a shared responsibility, and RadMail is a tool, not a guarantee of compliance. RadMail signs a BAA and processes on BAA-covered, encrypted, audited infrastructure, with a pluggable per-org regime (SOC 2, HIPAA, FedRAMP-CUI, BSA-AML) assembled from PolicyPacks. The SOC 2, HIPAA, and FedRAMP/CUI packs are architecture and roadmap, not certifications we hold. For regulated tenants, HIPAA handling is draft/ingest-only at this stage. Using RadMail does not by itself make your organization compliant.