Self-hosted email and groupware

Running your own mail, calendar and contacts is achievable with open-source servers such as Stalwart or mailcow, but deliverability, archiving and 24/7 operations decide whether it succeeds. This page explains the options, the architecture and the work involved.

On this page

When self-hosted email makes sense and which open-source stack fits

Self-hosted email makes sense when your organization needs direct control over where mail is stored, how long it is retained, and who can reach it, and when your team can carry the operating load of mail delivery, security patching and monitoring. For regulated organizations, the trade-off is usually between that control and the day-to-day work that Microsoft 365 or Google Workspace takes off your hands. Two open-source stacks cover most of the evaluation: Stalwart, an all-in-one mail and collaboration server, and mailcow-dockerized, a Docker-based bundle of mail components. Check the current versions at evaluation time, since the mailcow release 2026-09a and Stalwart v0.16.26 are the baselines this page was prepared against.

Stalwart serves IMAP, JMAP, SMTP, CalDAV, CardDAV and WebDAV from one server, so mail, calendar and contacts share one identity store, one set of TLS settings and one place to manage retention. It is licensed under *AGPL-3.0*, with an optional Enterprise License from Stalwart Labs, which matters if you plan to modify the code or offer it to third parties. mailcow-dockerized is licensed under GPL-3.0 and assembles Postfix, Dovecot, SOGo, rspamd and related components into a Compose-based stack that is quicker to stand up and easier for a small platform team to understand. It fits teams that want conventional, well-documented components. Stalwart fits teams that want a single modern server and are comfortable with a newer codebase. Confirm protocol coverage, clustering and archiving features against the release you select, because they change between versions.

Whichever stack you choose, decide early who will own sending reliability. Publishing SPF, DKIM and DMARC records, managing DKIM keys, watching reputation with major mailbox providers and handling bounce processing are ongoing tasks, not one-time setup. Ask your implementation partner to show the DNS records, key rotation procedure and monitoring plan before you commit. Also confirm where journaling or archive copies will be written, because retention obligations such as those in 17 CFR 240.17a-4 or the audit controls in 45 CFR 164.312 depend on that design, not on the mail server alone.

Requirements to agree before choosing a mail platform

Mail platforms fail most often because the requirements were never written down, not because the software was wrong. The cards below cover the questions that set the platform, the hosting model, the staffing and the migration plan. Agree on them with legal, security and operations before you compare Stalwart, mailcow or any other option.

Mailbox count and growth

Count active mailboxes, shared mailboxes, resource calendars and distribution lists, then estimate growth over the next three years. Include service accounts that send mail from applications, since they often appear late. The count drives storage sizing, backup windows and whether a single server or a clustered design is needed. Ask who owns mailbox provisioning and offboarding, because orphaned mailboxes are a common source of stored data that nobody is responsible for.

Client and protocol needs

List every client in use, including desktop Outlook, Apple Mail, Thunderbird, mobile devices and line-of-business software. Confirm whether each needs IMAP, JMAP, SMTP submission, CalDAV, CardDAV or ActiveSync, because support differs by platform and some clients only work with a subset. Test calendar sharing and delegation early, since these are the features users notice first when they are missing.

Sending volume and mail flows

Separate person-to-person mail from application notifications, newsletters and marketing messages. Each stream has different reputation, rate and bounce handling needs, and sending bulk mail from the same IP pool as staff mail can damage deliverability for both. Document which applications relay through the mail server, and confirm whether any of them need a dedicated sending path or a separate domain.

Domains and sender authentication

List every domain that receives or sends mail, including legacy and subsidiary domains. Each one needs SPF, DKIM and DMARC records that match how it actually sends. Decide who controls DNS, who rotates DKIM keys and who reads DMARC aggregate reports. Without a named owner, a DMARC policy set to reject can block legitimate mail after a change to a sending service.

Retention duties and legal hold

Identify which mail is subject to retention rules. For example, SEC Rule 17a-4(b)(4) requires broker-dealers to preserve communications relating to the business for not less than three years. Confirm with counsel whether journaling, a separate archive or immutable storage is required, and how a legal hold suspends deletion for named custodians. Ask how deleted items, retention periods and holds will be evidenced during an examination.

Data residency and transport security

Record where mailboxes, backups, logs and search indexes may physically reside, and whether any sub-processor or relay outside your control touches the message path. Mail in transit should be encrypted, and the HIPAA Security Rule requires technical safeguards against unauthorized access to ePHI sent over electronic networks, along with audit controls for systems holding ePHI. Confirm whether those obligations apply to your mail before you set the scope.

Availability targets and on-call ownership

Set the acceptable outage window for mail, the recovery point for mailbox data and who responds at night and on weekends. Mail is often the first system users notice when it fails, and inbound messages are usually retried by sending servers for a limited period, which makes queue monitoring important. Decide whether your team, a partner or a managed provider holds the pager, and write that split into the contract before go-live.

Stalwart, mailcow and other mail options compared

This table compares Stalwart and mailcow against each other and against hosted Microsoft 365 or Google Workspace on protocol coverage, licensing, architecture, administration, fit by organization size and operating effort. Use it to shortlist options, then confirm feature support against current vendor documentation rather than assuming parity between platforms.

DimensionStalwartmailcow
Protocol coverageIMAP, JMAP, SMTP, CalDAV, CardDAV and WebDAV in one server; confirm current support details in vendor documentationContainerized stack built around Postfix, Dovecot and SOGo for mail, calendar and contacts; confirm the exact protocol and client set for your version
LicensingAGPL-3.0 with an optional Enterprise License from Stalwart Labs; review obligations with counsel before modifying codeGPL-3.0; confirm obligations for any modifications and bundled components with counsel
ArchitectureSingle all-in-one server binary with its own storage and directory layer; fewer moving parts to operateDocker Compose based multi-container stack; components can be inspected and replaced individually, at the cost of more containers to maintain
Administration modelConfiguration and administration through the vendor's tooling; confirm the admin interface and automation options for your environmentWeb administration console with DNS and certificate helpers; familiar to operators who already run Docker-based services
Fit by organization sizeOften evaluated by teams comfortable with newer server software and who want a consolidated stackOften evaluated by small to mid-sized teams that want a well-documented, widely deployed stack
Operating effortLower component count, but fewer community runbooks; plan for internal skills or a partnerMore components to patch and monitor, with a large body of community documentation
Hosted Microsoft 365 or Google Workspace (baseline)Vendor operates infrastructure, deliverability and compliance tooling; you pay per user and accept vendor-controlled hostingVendor operates infrastructure; you keep fewer knobs on deliverability and data location but spend less on operations

Reference architecture components and how they connect

A self-hosted mail platform is a chain of dependencies, and most failures start at a link the mail server does not control. The components below map to the places where messages are accepted, authenticated, filtered, stored and retained, so each one can be validated on its own before go-live.

Mail server and protocol endpoints

This is the core service, either Stalwart (which serves IMAP, JMAP, SMTP, CalDAV, CardDAV and WebDAV in one binary) or the Postfix, Dovecot and SOGo stack that mailcow bundles. Confirm which protocols each client group needs, including calendar and contact sync on phones, before choosing. Validate with real client tests for send, receive, folder sync and free/busy lookups, not only with a successful SMTP handshake.

DNS authority and sender records

Your MX, SPF, DKIM and DMARC records live in DNS you control or delegate, and receiving servers read them before they read your message. A wrong MX target or a missing DKIM selector causes silent rejection or spam placement, so treat zone changes as versioned, reviewed changes. Validate each record with an external checker and a test message to a major provider, and confirm who can edit the zone during an incident.

TLS certificate management

Every client and every peer server negotiates TLS with the mail host, so an expired certificate stops mail in both directions and can also break mobile sync without an obvious error. Decide whether issuance is automated through an ACME client or handled by an internal certificate authority, and set renewal alerts well before expiry. Validate the full chain and the certificate names used by MX, submission and IMAP endpoints.

Outbound relay or smart host

Many self-hosted deployments send through a relay because residential or cloud IP ranges often have poor reputation or block outbound port 25. A relay changes the sending IP that DMARC and SPF evaluate, so SPF must authorize it and reverse DNS must match the sending host. Validate by confirming the path in message headers and checking that bounce handling and feedback loops reach an address you read.

Filtering and abuse layer

Inbound filtering for spam, malware and phishing sits between the internet and mailboxes, and it decides what users never see. Confirm how quarantine release works, who receives false-positive reports, and how DMARC aggregate reports are parsed. Validate with test samples for each category and by reviewing the quarantine queue on a schedule, since a filter that drops legitimate mail is harder to detect than one that lets spam through.

Storage, backup and archive store

Mailboxes, the archive store and the backup set are separate things with separate retention rules. A backup proves recovery, while an archive proves what was sent and received, and regulated sectors such as broker-dealers and HIPAA-covered entities need the second as well. Confirm which records the SEC retention rule or HIPAA audit controls apply to your organization, then test a full mailbox restore and an archive search before go-live.

Identity directory and monitoring

Mailbox logins can run against a directory such as LDAP or an OIDC provider where the chosen platform supports it, and that dependency means a directory outage also locks users out of mail. Pair the directory with log collection and alerting on queue depth, authentication failures, TLS expiry and DNS drift, so a silent break shows up before users report it. Confirm which SSO methods your chosen server supports before designing around them.

How SPF, DKIM and DMARC decide whether your mail arrives

Sender authentication tells receiving servers whether a message claims to come from a domain that the domain owner actually authorized. SPF publishes the list of servers allowed to send for your domain, DKIM attaches a cryptographic signature that the receiver checks against a public key in DNS, and DMARC tells receivers what to do when SPF or DKIM fails and asks them to send you reports. Each record is only as good as the mail flow behind it, so the work is to inventory every system that sends as your domain (mailboxes, application notifications, marketing tools, scanners, helpdesk platforms) before you publish a restrictive policy.

  • SPF scope: Keep the SPF record to a single TXT record per domain and stay within the DNS lookup limit. Every third-party sender you forget becomes a false failure once you enforce.
  • DKIM key management: Generate keys per domain and selector, store private keys with restricted file permissions, and rotate on a schedule your change process can support. Publish the new selector before retiring the old one so in-flight mail still verifies.
  • DMARC progression: Start with p=none and a rua address to collect aggregate reports for several weeks, then move to p=quarantine and finally p=reject once legitimate sources pass. Use pct only as a temporary control during the move.
  • Alignment: DMARC needs the visible From domain to align with the SPF or DKIM domain. Check the envelope and header domains in your mailcow or Stalwart configuration, not only the DNS records.
  • Reporting operations: Someone must read aggregate reports and act on them. Assign that owner before enforcement, because unreviewed reports hide the sender you broke.

Confirm with your mailcow or Stalwart deployment exactly which selectors and sending hosts you run, and with each third-party sender the records they require.

Build in-house, use a partner or buy Microsoft 365 or Google Workspace

This table compares running open-source mail in your own team, running it with an implementation partner, and buying a commercial suite on control, data location, effort, skills and cost structure. Read each row as a question to answer for your own mail estate, and note that the cost row describes who is paid for what rather than amounts. The main takeaway is that open-source mail moves cost from licenses into staffing, change control and evidence work, so the right model depends on the skills you already have.

FactorIn-house open sourceOpen source with a partnerCommercial SaaS suite
Control over mail data and routingFull control over mailbox storage, transport rules and retention settingsFull control retained by you, with the partner configuring and documenting it inside your environmentControl limited to vendor-provided admin settings and tenant options
Data locationWherever you host the servers, including your data center or private cloudWherever you agree in the written scope, typically your own infrastructureVendor-selected regions, with data residency options to confirm with the vendor
Skills requiredMail transport, DNS, TLS, anti-spam tuning, backup and restore, and on-call responseYour staff own operations, with partner skills used for design, hardening and handoverTenant administration, licensing and vendor support; little mail infrastructure work
Effort to run day to dayOngoing patching, queue monitoring, blocklist checks and user account administrationShared between your team and the partner under the agreed support modelMostly vendor-operated, with your team handling accounts, policies and access reviews
Deliverability ownershipYour team maintains SPF, DKIM and DMARC records and monitors reputationPartner sets up and documents the records, and your team keeps them currentVendor manages sending infrastructure, while you still own your domain records
Archiving and retention evidenceYou design and prove retention, including any broker-dealer or health records rules that applyPartner can build and document retention and export procedures for your assessorDepends on licensed archiving features, which you should confirm against your obligations
One-time costsDesign, build, migration tooling, testing and documentationScoped implementation, migration, hardening and knowledge transfer, quoted from written scopeMigration services, tenant setup, and any configuration or data cleanup work
Recurring costsStaff time, hosting, storage, backups, monitoring and security updatesSupport retainer or staff time, hosting, storage, backups and periodic upgradesPer-user subscriptions and add-on licenses, typically billed on a recurring schedule
Main cost driversMailbox volume, storage growth, availability targets and staff skill gapsScope of migration, number of domains, support hours and compliance evidence depthUser count, licensed feature tier, archiving add-ons and data volume

How a self-hosted mail rollout runs, phase by phase

The sequence moves a mail platform from business requirements to a tested, documented service that your own team or a partner can operate. Each phase closes with a written output that gates the next, so cutover happens only after mail flow has been proven.

  1. Define requirements and scope the platform

    The client's mail, calendar and contacts owners, security lead and compliance lead agree on the mailbox count, domains, protocol needs (IMAP, JMAP, SMTP, CalDAV, CardDAV), mobile and desktop clients, retention duties and hosting location. The team then selects between Stalwart and mailcow against those requirements and records the reasons. The phase closes with a signed requirements document and a fixed-scope statement of work that lists what is in and out, including who runs operations after handover.

  2. Design DNS, routing and identity

    Engineers design the DNS records for MX, SPF, DKIM and DMARC, plus autodiscovery and TLS certificates, and map how mail flows in and out, including any third-party senders such as marketing tools, scanners and line-of-business applications. Identity integration, admin roles and the outbound relay or IP reputation approach are decided here. The output is a reviewed DNS and routing design with a change plan, and the client's DNS owner confirms that every record can be published and rolled back.

  3. Build the lab environment

    The platform is installed in a non-production environment inside the client's own cloud account, data center or private network, using the agreed configuration and hardening baseline. Backups, monitoring, logging and the admin access model are set up at this stage so they are tested before any user data arrives. The phase ends with a lab build record covering versions, configuration, backup restore tests and a list of open defects.

  4. Run a pilot with a small group

    A small group of users, chosen by the business owner and typically including people from IT and the compliance team, moves to the new platform on a test domain or a limited set of addresses. They test sending and receiving with external providers, calendar invitations, shared mailboxes, phone and desktop clients, and search. Feedback is logged against the requirements. The pilot is complete when agreed acceptance criteria are met and open issues are either fixed or accepted in writing.

  5. Test coexistence and mail flow

    While the existing Microsoft 365 or Google Workspace tenant stays live, engineers test mail routing between the two systems, shared calendar free and busy lookups, and delivery to external recipients. They measure deliverability with seed addresses and check that DMARC reports arrive and that outbound mail passes SPF and DKIM checks. This phase confirms that nothing is lost or duplicated during the transition. The output is a signed mail flow test report that the cutover plan depends on.

  6. Migrate mailboxes in stages

    Mailboxes, calendars and contacts move in batches sized to the client's change tolerance, with a verification step after each batch that compares item counts and sample content. Migration tooling, throttling and any source-side limits are confirmed with the existing provider before the run. Users keep working during the batch window where possible. The phase closes when each batch passes its verification log and the migration exception list is cleared or accepted.

  7. Cut over and decommission the old tenant

    The MX records are switched, outbound routing is updated, and the team watches delivery, bounce rates and DMARC reports through the change window with a documented rollback path. Once the new platform is stable, the old tenant is decommissioned on a schedule that respects archive and retention obligations, including any export of existing mail that must be kept. The output is a cutover record and a decommissioning certificate listing what was retired, exported and retained.

  8. Hand over operations and runbooks

    The platform moves into steady-state operations, whether run by the client's team, by DeployOpen under a retainer, or by a mix of both. Runbooks cover user onboarding, certificate and DKIM key rotation, backup restore tests, upgrade procedures, incident response and archive exports. Responsibilities are assigned in a RACI table that names who handles each task. The handover is complete when the operating team has performed a restore test and an upgrade rehearsal on their own.

Migrating mailboxes, calendars and contacts from Microsoft 365 or Google Workspace

The migration plan should be built around three data types that move differently: mailboxes, calendars and contacts. Mailbox content is usually copied over IMAP from the source tenant while the old service stays live, which lets you run a first full sync and then delta passes until cutover. Calendars and contacts are harder because they live in CalDAV and CardDAV collections and in provider-specific formats, so you should confirm with your migration tool exactly which item types, recurrence rules, attachments, room resources and shared calendars it carries across. Stalwart and mailcow both expose these protocols, but the source side (Microsoft 365 or Google Workspace) determines what can be extracted cleanly. Ask your implementation partner for a written mapping of every data type and the known exceptions before any mailbox is touched.

Coexistence is the period when some users are on the new server and others are still on the old one, and mail must route correctly in both directions. Plan the MX and routing changes so that inbound mail for migrated and unmigrated users reaches the right place, and decide how the old tenant forwards or connects to the new domain. Shared mailboxes and distribution lists need separate decisions: confirm who owns each one, which members receive copies, whether send-as and send-on-behalf rights survive, and whether group membership is recreated manually. Some items will not move cleanly, including provider-specific rules that depend on cloud features, certain delegation models, and historical message state. Document these gaps for each user group so that business owners can accept them or plan a manual fix.

Before cutover, run a pilot group through the full path and test client reconnection on desktop and mobile clients, since passwords, app-specific credentials and certificate trust often need reconfiguration. Send and accept calendar invitations from external organizations and from colleagues still on the old platform, and confirm that free and busy lookups and meeting updates behave as expected. Check that contacts sync and that migrated mail appears in the correct folders. Treat any failed case as a blocking item for that wave. Rollback steps for DNS, routing and client settings should be written and rehearsed, and the decision to proceed should be recorded with the owners of each affected group.

Archiving and retention obligations for self-hosted email

Does SEC Rule 17a-4 apply to email we host ourselves?

It applies if you are a broker-dealer subject to the rule, and self-hosting does not change that. SEC Rule 17a-4(b)(4) requires broker-dealers to preserve originals of communications received and copies of communications sent relating to the business, including inter-office communications, for not less than three years. The rule is written around the records, not the server, so your mail platform, archive and retention settings must together keep those communications for the required period. Confirm with your compliance counsel which communications are in scope, how long each category must be kept, and whether your regulator expects a specific storage format. DeployOpen can build and document the archive and retention controls you decide on, but the determination of your obligations stays with your organization and counsel.

What does HIPAA require for email containing ePHI?

The HIPAA Security Rule expects technical safeguards for electronic protected health information sent over networks and audit controls for systems that hold it. Under 45 CFR 164.312(e)(1), you must guard against unauthorized access to ePHI transmitted over an electronic communications network, which in practice means enforcing TLS for SMTP and client connections and reviewing that enforcement. Under 164.312(b), you must implement audit controls that record and examine activity in systems containing ePHI, so mailbox access, administrative changes and archive searches need logging that someone actually reviews. Whether a given email contains ePHI, and how your risk analysis treats it, is your decision with your privacy and security officers. DeployOpen configures and evidences these controls; it does not certify HIPAA compliance.

What is journaling and how does it differ from mailbox backup?

Journaling copies every message that passes through the mail system, at delivery or submission, into a separate archive store that users cannot edit or delete from their mailboxes. A backup is a point-in-time copy used to restore systems after failure, and it rarely gives you a complete, searchable record of communications over years. For retention obligations, you usually need journaling or an equivalent capture point, plus immutable storage with defined retention periods. With Stalwart or mailcow, confirm which capture mechanisms the version you deploy supports, how journal copies are formatted, and whether the archive can be exported in a format your reviewers and auditors accept. Test a restore and a search before you rely on the archive.

How do legal hold and e-discovery work on self-hosted mail?

A legal hold suspends deletion for specific custodians or messages while a dispute or investigation is open, and e-discovery is the process of identifying, collecting and producing relevant messages. Both depend on the archive being searchable by custodian, date range, sender, recipient and keyword, and on holds being applied independently of users' own deletion actions. Your legal team sets the hold scope and the production format. Engineering needs to confirm that holds override retention cleanup, that hold actions are logged, and that exports preserve metadata. Ask your implementation partner to demonstrate a hold on a test mailbox, an export and a log entry for each action, and agree the evidence set in writing before go-live.

Who decides our retention periods for self-hosted email?

Your organization decides retention periods, with legal, compliance and records management advising on them. Regulations such as SEC Rule 17a-4 and HIPAA set requirements for particular categories of records and safeguards, but they do not give one universal period for every email, and they interact with state law, contracts and internal policy. The platform implements the schedule you approve, for example by tagging categories, setting retention rules per domain or group, and applying deletion only after holds clear. DeployOpen can configure and document the schedule as written by your organization and prepare evidence that it runs as specified. It does not interpret the law or give legal advice.

What evidence will auditors expect from our email archive?

Auditors typically ask for the written retention policy, the configuration that enforces it, samples showing messages were captured and kept for the required period, records of holds and deletions, access logs for the archive, and proof that backups of the archive can be restored. For regulated email, expect requests to show who can administer the archive and how those privileges are reviewed. Build an evidence register early: each control, its owner, where the evidence is stored and how often it is produced. DeployOpen prepares controls and evidence for compliance assessments. Independent auditors and certification bodies issue any opinion or attestation, and DeployOpen does not issue those.

What does archiving cost for self-hosted email?

Cost has three main parts: the people who design and operate the archive, the storage that holds years of messages, and the effort to run holds, exports and audits. Storage grows with mailbox volume, attachment size and retention length, and it recurs every year, so model it from your message volumes and retention schedule rather than from a generic figure. Implementation effort is one-time, and ongoing administration, log review and restore testing recur. When requesting a quote, ask the partner to separate design, configuration, storage sizing, evidence preparation and continuing support, and to state which items are fixed-scope and which are retainers. DeployOpen agrees scope and deliverables in writing before work starts and does not publish a price list.

Can DeployOpen run our email archive compliance for us?

DeployOpen can plan, implement, configure, harden and document the archive, retention, journaling and audit logging controls in your own infrastructure, and it can prepare the evidence your assessors request. It does not issue audit reports, attestations or certifications, and it does not decide your legal obligations. Your compliance and legal functions own the retention schedule and hold decisions. Ongoing operation, log review and periodic restore tests can be covered by a retainer if your team wants that support. Put the division of responsibilities into the written scope so that every control has a named owner.

Systems the mail platform must connect to and what each involves

A self-hosted mail platform is only as good as the systems around it. Identity, DNS, logging, backup and archiving each take a share of the work, and each has its own owner. Read the cards to see which connections your team must build, which it must confirm, and where the risk sits.

Identity provider and SSO

Authentication decides who can log in to mailboxes, calendars and contacts, and how quickly access is removed when someone leaves. Confirm which protocols your chosen server supports for single sign-on, and whether mail clients will still authenticate through a password or app credential when the identity provider is down. Test offboarding end to end, including sessions held by phones and desktop clients.

Directory for user and group sync

Mailbox creation, distribution lists and shared calendars depend on accurate user and group data from a directory. Decide whether accounts are provisioned by sync from the directory or created by hand, and who owns each attribute. Unsynced groups cause misdelivered mail and broken free and busy lookups, so test group membership changes before go-live.

DNS provider and mail records

Your DNS provider hosts the records that receiving servers use to trust your mail: SPF, DKIM and DMARC, along with MX and autodiscovery entries. Changes often take effect slowly, and a wrong record can stop outbound mail or cause legitimate messages to be rejected. Confirm who can edit zone records, how DKIM keys are rotated, and how DMARC reports will be read.

SIEM and log forwarding

Mail servers generate authentication, delivery and administrative logs that security teams need to investigate phishing, account takeover and data loss. Forward logs from the mail server, the reverse proxy and the admin console to your SIEM, and confirm they keep timestamps consistent. Define which events trigger alerts, because a full log feed without tuning becomes noise.

Backup platform

Mailboxes, indexes and configuration must be restored together, or users may find a mailbox that no longer matches its domain settings. Confirm that your backup tool captures mail storage and the database, and that restores are tested into a separate environment. Ask who holds the encryption keys and how long backups are kept.

Archive and compliance store

Where a regulator requires retention of business communications, the mail platform alone is rarely enough. SEC Rule 17a-4(b)(4) requires broker-dealers to preserve originals of communications received and copies of those sent for not less than three years, and HIPAA's audit control requirement under 45 CFR 164.312(b) calls for records of activity in systems that hold ePHI. Decide whether the archive is a separate store with its own immutability controls, and confirm it with your compliance team.

Calendar, contact and file sharing clients

Users reach mail through desktop, mobile and web clients, and each client handles CalDAV, CardDAV, IMAP and JMAP with its own quirks. Calendar invites and shared files also link to file sharing and groupware tools, so links that break at the boundary generate support tickets. Test the client list you actually support, and document which features (such as free and busy lookup or push) work on each.

Compliance evidence that self-hosted email supports

This table maps common email and groupware control areas to the evidence an assessor typically asks for and where self-hosted deployments tend to fall short. Use it to plan what your team must produce, since auditors and certification bodies, not DeployOpen or the mail platform, issue the attestations. The main takeaway is that retention, audit logging and transport encryption are configured and evidenced by you, so they need owners before the first assessment.

Control areaWhat the assessor expects to seeWhere self-hosted mail often falls short
Retention for regulated communications (SEC Rule 17a-4(b)(4))Documented retention period of at least three years for business communications, with originals and sent copies preserved and retrievableMailboxes purged by user deletion or default quotas, with no archive or journaling policy in place
Transmission security (HIPAA 164.312(e)(1))TLS enforced for SMTP, IMAP and web access, with configuration evidence and certificate management recordsOpportunistic TLS accepted without enforcement, or expired certificates on submission and client ports
Audit controls (HIPAA 164.312(b))Logs of authentication, administrative changes and message access, retained and reviewed on a defined scheduleLogs rotated too quickly, stored on the mail host only, or never reviewed
Access control and administrator privilegesNamed admin accounts, MFA, role separation and periodic access reviews tied to the identity providerShared administrator logins, legacy IMAP basic authentication left enabled, and stale accounts
Sender authentication (SPF, DKIM, DMARC)Published records and DMARC reports showing enforcement status and monitored alignmentDMARC left at monitor-only indefinitely, or DKIM keys never rotated and unrecorded
Backup and restoreBackup schedule, encryption status, restore test results and recovery time recordsBackups that exist but have never been restored in a test, or mail stores excluded from backup scope
Change and vendor managementVersioned configuration, patch records for the mail server and its containers, and documented upgrade approvalsUpgrades applied ad hoc with no record of version, date or approver
Encryption at restDocumented storage encryption for mail data and backups, with key custody and access rulesMail stores on unencrypted volumes or keys held by the same administrators who access data

Who handles operations once the mail server is live

Mail operations are a continuous workload, not a one-time install. Once Stalwart or mailcow is live, someone must own patching and release tracking, certificate renewal for every hostname that clients and remote servers use, blocklist monitoring for the sending IP addresses, queue and spam filter tuning, backup restore tests, capacity planning for mailbox growth, incident response, and on-call coverage for outages that affect delivery. Both projects publish releases on their own schedules. Stalwart's v0.16.26 (October 11, 2026) and mailcow's 2026-09a (October 6, 2026) are examples of the cadence to expect, so the first question for any operator is who reads release notes and who schedules the upgrade window. Confirm the upgrade path with the project before you commit to a date.

The three staffing models divide that workload differently. An internal team needs at least two people who can run the mail stack, understand SPF, DKIM and DMARC failures, and work through a message-queue incident at night. It also needs a documented escalation path and a restore procedure that has been tested, not assumed. A partner retainer moves patching, monitoring, and on-call to an external team under a written scope, with your administrators keeping ownership of mailboxes, domains, and retention decisions. A hybrid model keeps identity, mailbox administration and policy in-house while the partner handles the infrastructure layer. Each model requires the client to supply named owners, access to DNS and identity systems, and a decision on who may approve changes to sending policy.

When you request a quote, ask for the operational scope in writing: which release tracking and patch windows are included, how certificate renewal is automated and monitored, how blocklist checks are run and reported, what response times apply to outages, and whether restore tests are performed on a schedule with evidence kept. Because these terms drive recurring cost, ask which items are one-time setup and which recur each month. Confirm with your legal and records team whether the archive you run is the system of record for any communications subject to retention rules, since that decision changes how strict the on-call and backup obligations must be.

Self-hosted email questions buyers ask before hiring help

What drives the cost of self-hosting email with Stalwart or mailcow?

The cost splits into one-time work and recurring work, and the mix depends on the scope agreed in writing. The one-time items are design, DNS and authentication setup (SPF, DKIM and DMARC), the migration, and the archiving design. The recurring items are infrastructure, backups, patching, monitoring, mail queue and reputation watching, and support. Costs rise with mailbox count, the number of domains, the retention period, the volume of mail that needs to be indexed or archived, and whether you need high availability across sites. Costs fall when the scope is narrow, the source mailboxes are clean, and the team can handle first-line support. Ask every quote to separate setup fees from monthly operations, to state who owns the servers and licenses, and to list what is excluded, such as third-party archiving products or paid Stalwart Enterprise License terms. DeployOpen does not publish a price list, so the quote follows from the written scope.

How long does a migration from Microsoft 365 or Google Workspace take?

The timeline is set by data volume, number of mailboxes, the cutover model and how much history must arrive intact, not by the platform alone. The main drivers are total mailbox size, the number of shared mailboxes, calendars and contacts that must move with users, the complexity of routing and aliases, and how many domains and users change at once. A phased approach that moves pilot users first, then shared mailboxes, then the rest, usually adds planning time but reduces risk. Ask the partner for a written migration plan that covers pre-migration DNS lowering, sync windows, a rollback path, and a validation step for message counts and folder structure. No duration is given here because none has been researched for this page. Request a timeline that names the assumptions behind it, so you can test those assumptions against your own mailbox inventory.

Can a small IT team run a self-hosted mail server without a dedicated admin?

A small team can run it, but only if the operating model is defined before go-live. The work that tends to fall on staff includes watching the mail queue, responding to blocklist listings, renewing TLS certificates, applying platform updates such as mailcow release 2026-09a, checking backups, and handling user requests for aliases, quotas and mailbox restores. Those tasks are routine but time sensitive, and an outage or a blocklisting can stop business email for everyone. Decide who is on call, who approves changes, and who answers when delivery fails. Many organizations keep internal ownership of identity and user administration while contracting platform operations, monitoring and patching to a partner under a retainer. Ask for a responsibilities matrix that lists each recurring task, the owner, and the escalation path.

What does the client team need to provide for a self-hosted email project?

The project depends on several inputs that only your organization can supply. You need access to the DNS zones for every mail domain, the ability to create SPF, DKIM and DMARC records, and an owner who can approve changes to MX records. You also need a current inventory of mailboxes, shared mailboxes, distribution lists, aliases, and external systems that send mail, such as scanners, line-of-business applications and marketing tools. Identity inputs matter too: your directory source, password and MFA policy, and how user accounts are provisioned. Finally, you need a named business owner for retention rules and legal hold decisions, plus a hosting environment with the network and storage your security team has approved. Missing inputs are the most common reason a fixed-scope timeline slips, so ask the partner to list them in the statement of work.

How much deliverability risk comes with running my own mail server?

Deliverability risk is real in the sense that it is operational, not theoretical. Receiving servers judge a sender by its IP reputation, its domain alignment, and whether SPF, DKIM and DMARC pass and align. A new server starts with no history, so mail to large providers may be filtered until reputation builds. The safeguards to plan for are a correctly configured reverse DNS record, consistent SPF includes for every sending service, DKIM signing on all outbound paths, and a DMARC policy that moves from monitoring to enforcement only after reports show legitimate senders pass. Ask the partner how they will monitor DMARC aggregate reports, check blocklists, and handle a listing if one occurs. Also confirm whether your outbound IP is shared or dedicated, because that choice changes your exposure.

How do SEC and HIPAA retention rules change the email architecture?

Retention obligations change the design from the first day, because email must be captured, kept and retrievable, not just delivered. For broker-dealers, SEC Rule 17a-4(b)(4) requires preserving originals of received communications and copies of sent communications relating to the business, including inter-office communications, for not less than three years. That means you need a write-once or otherwise compliant archive, a defined retention schedule, and a way to export records on request. For healthcare, the HIPAA Security Rule at 45 CFR 164.312(e)(1) requires technical measures that guard against unauthorized access to ePHI sent over networks, and 164.312(b) requires audit controls that record and examine activity in systems holding ePHI. In practice, you need encrypted transport, mailbox access logging and retention of those logs. Your compliance counsel decides whether a given rule applies and how long to keep records. Ask the partner to show how archiving, journaling, encryption and audit logs map to each requirement, and to name what it will not cover.

Does a self-hosted mail platform support all the protocols my clients need?

Protocol coverage depends on the platform you choose and on how it is configured, so confirm it against your own client list. Stalwart's project documentation lists IMAP, JMAP, SMTP, CalDAV, CardDAV and WebDAV, which covers mail, calendar, contacts and file access in one server. mailcow is built around a different stack of components, so check its documented support for each client type before you commit. Test the clients you actually use, including older desktop mail applications, mobile devices, calendar sharing, and delegated access. Also verify how push notifications and shared calendars behave, since these are the features users notice first when they are missing. Ask the partner for a client compatibility test plan and results before the pilot starts.

How DeployOpen scopes, builds, hands over and supports self-hosted email

DeployOpen plans, implements, hardens, documents and supports self-hosted open-source platforms inside your own cloud accounts, data centers or air-gapped networks. For email and groupware, that means agreeing a written scope for Stalwart or mailcow before any work starts, then delivering the mail, calendar and contact services with the evidence your compliance team needs.

How we scope

  • Inventory mailboxes, domains, shared mailboxes, calendars and contacts that must move or stay on the current platform.
  • Confirm sending sources, DNS ownership and the SPF, DKIM and DMARC records each domain will publish.
  • Define which protocols clients need, such as IMAP, JMAP, SMTP, CalDAV and CardDAV, and where they connect from.
  • Record archiving and retention obligations with your legal and compliance owners before choosing a retention design.

How we build

  • Deploy the chosen platform in your infrastructure, with mail flow, TLS configuration and access controls set to your written standards.
  • Test inbound and outbound delivery, authentication results and client compatibility across the protocols in scope.
  • Configure audit logging and retention settings so that message records can be reviewed against your stated obligations.
  • Integrate directory sign-in, alerting and backup jobs, and validate restores of mailboxes before cutover.

How we hand over

  • Deliver architecture, configuration and runbook documentation covering mail flow, DNS records and mailbox administration.
  • Provide migration records, cutover checklists and a list of known gaps agreed with your team before go-live.
  • Walk operations staff through monitoring, queue handling, certificate renewal and user onboarding procedures.
  • Produce compliance evidence packs that show controls in place, for your auditors to assess independently.

How we support

  • Ongoing support, upgrades and continuous testing are available as retainers once the platform is live.
  • Plan and document platform upgrades and security patches within the agreed support scope.
  • Confirm with us during scoping which operational tasks stay with your staff and which move to a retainer.

Talk to our team.

Tell us what you're working on, whether it's a deployment, an audit, a security test or a cyber range. You'll speak with an engineer who can help you scope it.

  • 30-minute call: free, with no obligation.
  • NDA on request: we can sign before you share details.
  • Clear next steps: a scope and plan after the call.