Host community certificate Wi-Fi
A community location can give eligible members a way to connect with their own device certificates instead of sharing a Wi-Fi password. Hosting means operating an access point and an isolated network, looking after the location, and working with the people responsible for identity and authentication.
Choose the operating model
Section titled “Choose the operating model”Approved IrregularNet location
Section titled “Approved IrregularNet location”- Identity and certificates: IrregularNet’s platform would manage enrollment, the certificate authority and RADIUS authentication.
- Member experience: The goal is for one certificate per device to work at all approved locations.
- Local responsibilities: The approved host operates its access point, isolated member network, transport and onsite support.
- Availability: Additional locations are not yet available for self-service activation.
Independent organization network
Section titled “Independent organization network”- Identity and certificates: Your organization owns its identity provider, enrollment, certificate authority and RADIUS policy.
- Member experience: Your users enroll under your own realm and policy.
- Local responsibilities: Your organization also owns issuance, offboarding, revocation, backups and recovery.
- Availability: Requires an independently planned and operated deployment.
The two models do not automatically federate. The current IrregularNet code and configuration include Muse pilot assumptions; they are not a turnkey independent realm. For a separate deployment, use the organization certificate Wi-Fi guide as the technical starting point and assign your own operators and recovery authority.
For the decisions needed to adapt this codebase, use the independent-realm implementation plan and its downloadable realm worksheet. It maps the current Muse-specific policy to the installer milestones.
A realm is the membership and certificate trust domain. Several approved locations can belong to the same realm. An independently owned realm has its own trust and admission policy; using the same SSID does not join the realms.
Set up with Codex, Claude or another AI assistant
Section titled “Set up with Codex, Claude or another AI assistant”Download the IrregularNet hosting skill (SKILL.md).
Attach the file to your assistant, or paste its contents, and give it your exact
router/access-point model and firmware. The skill checks the current roadmap,
operating model and approval requirements, then helps prepare an equipment-specific
configuration preview, recovery plan and onsite acceptance checks. It also covers
planning a separate realm.
For example:
Read the attached IrregularNet hosting skill. I want to host a shared community location using [my AP/router model and firmware]. Start with a read-only assessment and prepare a concrete configuration preview.
Prospective hosts can use it to prepare a site plan while approval or provisioning is pending. Keep site secrets, private keys and downloaded device profiles out of the conversation; the assistant should use approved private references when those inputs are needed.
Equipment options
Section titled “Equipment options”Start by checking equipment you already own. These are configuration paths to evaluate, not an IrregularNet compatibility list or a buying recommendation. Confirm the exact model, firmware, Enterprise mode, VLAN behavior and private authentication transport before changing or purchasing equipment.
- Controller-managed access points: use an external RADIUS profile and a separate member network. Ubiquiti documents Enterprise Wi-Fi settings and RADIUS configuration. Controller features alone do not install IrregularNet’s certificate policy.
- Supported OpenWrt equipment: evaluate Enterprise authentication, the installed wireless package and VLAN/firewall isolation for that release. OpenWrt’s encryption guide describes the external-RADIUS configuration. A supported hardware model alone does not prove this complete flow works.
- An independent managed fleet or self-hosted realm: choose identity, certificate issuance, device management and RADIUS together. The organization deployment guide compares the implementation paths and their operating responsibilities.
Use documentation matching your installed release. Do not copy legacy WPA, TKIP, password-only or captive-portal examples into the certificate network.
Plan the equipment and connection
Section titled “Plan the equipment and connection”A shared password-only router cannot provide this certificate-based service. Check the documentation for the exact access point, controller and firmware you intend to use. Your equipment needs an enterprise 802.1X Wi-Fi mode that can send EAP-TLS authentication to a RADIUS server. Confirm that it supports the network separation, management and recovery features your location needs. Existing equipment may work; capacity, roaming and supported client devices need testing at the site.
For a future approved location, plan for:
- A separate member VLAN or equivalent isolated network, with explicit DHCP, DNS and firewall policy. Check both IPv4 and IPv6, including paths to local management, staff and other private networks.
- A private, approved route to the authentication service. The service must identify the actual RADIUS client’s exact transport source and approved site binding; client-supplied NAS names and outer usernames are not site identity. Independent locations behind the same public NAT cannot safely be distinguished by that shared source alone.
- A distinct site secret delivered through an approved private process. Treat it as a password: do not put it in this worksheet, chat, tickets or public configuration examples. Never copy another location’s secret or CA keys.
- Protected administrator access and a separate way to reach the equipment if the new network or authentication service fails. Preserve the existing staff and guest networks until a tested change is approved.
- An onsite person with an approved test device, time to perform the checks, and authority to carry out the rollback.
The future multi-location provisioning process still needs to bind each site’s private source, secret and loaded RADIUS configuration together. A healthy access point or an authentication response alone does not prove that binding, network isolation or a working connection.
What a host would be responsible for
Section titled “What a host would be responsible for”An approved host would maintain its equipment, software updates, local network isolation and a way to contact its responsible operator. The host should agree on hours, coverage, capacity, incident handling and a privacy-conscious support process before offering access. Do not collect private keys, downloaded profiles or unnecessary member information for troubleshooting.
The planned location-operator workflow would let an assigned operator flag a member or restrict access at that operator’s location, with a required comment. It is not yet a live console. A location restriction is distinct from infrastructure administrators revoking a certificate across the service. Neither a restriction nor certificate revocation should be described as an instant disconnect: an existing session may continue until reauthentication or a separately verified disconnect.
Plan, test and activate
Section titled “Plan, test and activate”Once multi-location provisioning is released and the host has been approved, the expected sequence is:
- Approve and plan. Agree on the location, responsible operators, network boundary, support and rollback. Confirm who can approve activation.
- Stage. Create the location binding and private RADIUS transport through the reviewed provisioning process. Configure a separate test SSID and member network without replacing the existing service.
- Test onsite. Enroll an eligible device over an existing internet path, install its own profile, verify the expected server identity, and test a real connection and application. Check permitted access and denied access to staff and management resources over IPv4 and IPv6. Test an ineligible or revoked credential at fresh authentication, plus outage and rollback behavior. Keep the results and any unresolved device or firmware limitations.
- Activate after review. The host and platform operators review those results and approve the location. Monitor and support the service; repeat relevant checks after equipment or configuration changes.
These steps describe the intended release process, not actions available in the current pilot. An authentication acceptance alone does not prove the device joined, received the right network, or stayed connected. Revocation tests must measure what happens to already connected and resumed sessions rather than assuming they disconnect immediately.
Planning worksheet
Section titled “Planning worksheet”Download the plain-text planning worksheet to fill in locally, or use the checklist below. Downloading it does not submit an application.
Keep this as a local planning checklist. Record no passwords, site secrets, private keys, certificate profiles or member incident details here.
- Shared community location or independent realm
- Location and responsible host
- Primary and backup onsite operators
- Access point/controller and firmware to evaluate
- Expected coverage, devices and capacity
- Internet-sharing permission and intended opening hours
- Separate member network and IPv4/IPv6 isolation plan
- Proposed private authentication route and unique transport source
- Existing staff/guest service and alternate management path to preserve
- Onsite tester, test devices and available test window
- Who approves activation and who can perform rollback
- Support contact and outage/recovery plan
Discuss the plan privately with the Muse operator or your IrregularNet infrastructure contact. This page does not submit an application or create a location. Agree separately what location and contact information may be public.
Follow the roadmap
Section titled “Follow the roadmap”The implementation roadmap tracks live, prepared and planned work, owners and next acceptance gates. Repository access may require sign-in. Shared-realm admission and site controls are being prepared in MR !476. Independent-realm installation and cross-realm federation remain separate future work. Guide publication does not activate any of these capabilities.
For a member’s enrollment and device instructions, see the IrregularNet guide. For the broader work of running your own issuer and authentication service, see the organization deployment guide.