Agencies, freelancers, and in-house teams running more than one site at a time. Plus founders running a small portfolio (3–25 sites) who want one analytics fabric instead of one Plausible per project. The Agency plan unlocks multi-workspace, team membership with role-based access, the Event Explorer, exports, audit logs, and the MCP + Agent integrations.

How is multi-site organised?

Sites belong to workspaces; users belong to workspaces via membership. A user with active membership in workspace A sees A's sites; cross-workspace reads are impossible. Agency unlocks multiple workspaces per account so you can give each client their own fabric without sharing data across them.

Can a team member only see one client's sites?

Yes. Workspace membership is the access boundary. Add a team member to one workspace and they see only that workspace's sites, events, and attribution. RBAC roles (owner / admin / operator / analyst / support / readonly) refine what they can do inside that workspace.

Does AI visibility work per site or per workspace?

Per site, aggregated per workspace. The dashboard's overview rolls up the workspace's full site list by default; the FilterBar narrows to a single site. The Event Explorer + Attribution Engine drill into any one site or across all sites the operator has membership on.

How does the Claude/MCP integration work across many sites?

MCP is workspace-scoped. The Bearer key Claude uses is tied to a workspace; the read surface includes every site in that workspace. Ask Claude 'which sites are losing AI assistant traffic this week' and the MCP server returns the workspace-wide cross-section.

Do I need a workspace per client or can I run them all in one?

Up to you. Some agencies run one workspace per client for clean billing + audit boundaries; others run one workspace and add team members from the client side as readonly. Agency supports either model. For a custom infrastructure conversation, write to info@helperg.com.

Multi-site analytics

One analytics fabric for many sites — and the team that runs them

WebmasterID Agency is built for operators with more than one site to look after. Multi-workspace, team membership with RBAC, the Event Explorer, exports, audit logs, and Claude/MCP integration. AI-native by design, agency-ready by default.

Upgrade to Agency · $149 / mo Compare plans Talk to us

What teams get

Why teams with many sites benefit massively

Running multiple sites in 2026 used to mean running multiple analytics dashboards. With AI assistants entering the workflow, single-site tools start to feel limiting. WebmasterID is built for the portfolio shape.

Multi-workspace

Each client (or each project, each brand) gets its own workspace. Sites, events, attribution, and exports are workspace-scoped. Cross-workspace reads are impossible by construction — the access boundary is enforced at the SQL layer, not at the UI.

Unlimited team members

Add operators, analysts, and clients to the workspaces they need to be in. RBAC roles (owner / admin / operator / analyst / support / readonly) refine permissions inside each workspace. Suspended or revoked memberships lose visibility immediately.

Advanced attribution

Pro covers the basics; Agency unlocks recent attributed sessions, source-confidence labels, and deep filters across the workspace's full site list. Drill from a source to the underlying events via the Event Explorer.

Event Explorer

Browse the workspace's full event stream by date, traffic type, country, pathname, referrer, session, visitor, or campaign. Drill into any single event for the safe metadata view. Useful for incident investigation, debugging tracker installs, and confirming the system records the events you expect.

Exports

CSV + NDJSON exports of filtered event sets — for reporting, reconciliation, or piping into a client's own warehouse. No per-export gating on Agency.

Audit logs

Every admin action (member invited, role changed, repo mapping verified, etc.) is recorded in the workspace's append-only audit log. PII is sanitised; the trail is immutable; downstream compliance conversations get a real answer.

MCP + Agent integrations

The MCP server is workspace-scoped. Claude can read the workspace's full site list, attribution, and recommendations through one Bearer key. AI-assisted investigation across a portfolio is one prompt, not a tab-switching session.

Collaboration features

Invite team members by email — the invite token is single-use, SHA-256-hashed, expires in 7 days, and the email is sent via AWS SES. Roles are workspace-scoped. The legacy single-tenant owner pattern still works for solo operators.

Workflow

A day-in-the-life across many sites

What an agency operator actually does with WebmasterID Agency, written without hype.

Morning brief. Open the daily operator brief in the Agent surface (Agency plan). The brief lists the workspace's top actions for the day with noise-suppressed deduplication — duplicate findings across sites collapse into a single recommendation.
Ask Claude. Through MCP: "Which sites lost ChatGPT referral traffic this week?" Claude reads the workspace's attribution data and answers with the cross-section. The audit log records the read.
Drill into a specific site. FilterBar narrows to one site. Open the Event Explorer, filter by referrer source, drill into a single event for the safe metadata. Export the slice for the client report.
Prepare a code change. The Agent's recommendation is bound to the site's repo mapping. Ask Claude to read the recommendation, open the file, and propose a diff. You review; you ship.
Hand it back to the client. Invite the client as a readonly member of the workspace so they can verify the change took effect. Or export a CSV and send a digest. Same workspace, different access shape.

Privacy

Multi-site analytics that respects every site's audience

Privacy-first across the whole portfolio. The single-site posture extends to many sites without compromise.

No cookies, no fingerprinting
Every site uses first-party event recording with text/plain transport. Visitors carry no third-party identifier between sites — even if multiple sites in the workspace share an audience.
No raw IP storage
Country headers only, server-side. No IP-based cross-site linking, no fingerprint reuse across the workspace.
No data sale or sharing
Your workspace's data stays on infrastructure we own. We don't sell it, share it, or train shared models on it.
GDPR-aligned by design
See /privacy-policy and /data-processing for the full posture.

FAQ

Multi-site analytics, in plain English

Who is this for?: Agencies, freelancers, and in-house teams running more than one site at a time. Plus founders running a small portfolio (3–25 sites) who want one analytics fabric instead of one Plausible per project. The Agency plan unlocks multi-workspace, team membership with role-based access, the Event Explorer, exports, audit logs, and the MCP + Agent integrations.
How is multi-site organised?: Sites belong to workspaces; users belong to workspaces via membership. A user with active membership in workspace A sees A's sites; cross-workspace reads are impossible. Agency unlocks multiple workspaces per account so you can give each client their own fabric without sharing data across them.
Can a team member only see one client's sites?: Yes. Workspace membership is the access boundary. Add a team member to one workspace and they see only that workspace's sites, events, and attribution. RBAC roles (owner / admin / operator / analyst / support / readonly) refine what they can do inside that workspace.
Does AI visibility work per site or per workspace?: Per site, aggregated per workspace. The dashboard's overview rolls up the workspace's full site list by default; the FilterBar narrows to a single site. The Event Explorer + Attribution Engine drill into any one site or across all sites the operator has membership on.
How does the Claude/MCP integration work across many sites?: MCP is workspace-scoped. The Bearer key Claude uses is tied to a workspace; the read surface includes every site in that workspace. Ask Claude 'which sites are losing AI assistant traffic this week' and the MCP server returns the workspace-wide cross-section.
Do I need a workspace per client or can I run them all in one?: Up to you. Some agencies run one workspace per client for clean billing + audit boundaries; others run one workspace and add team members from the client side as readonly. Agency supports either model. For a custom infrastructure conversation, write to info@helperg.com.

More detail on the parts that matter for teams

Pricing

Agency at $149/mo unlocks multi-workspace, team, advanced attribution, Event Explorer, exports, audit logs, MCP + Agent.

AI-assisted analytics

The reading-partner shape — how Claude reads many sites' data through MCP without leaving the operator out of the loop.

Agent + MCP details

The technical surface Claude reads. Workspace-scoped, Bearer-authed, audit-logged.

AI visibility per site

Crawler intelligence + AI assistant referrals — what the workspace sees aggregated and per-site.

Attribution explainer

How directional attribution works and where the limits are.

Architecture

The data + auth + workspace model under the hood. Privacy-first by construction.

Agency analytics

The agency story laid out by use case — multi-client workflows, RBAC, audit logs.

Attribution analytics

Directional attribution across the portfolio with confidence labels per row.