Semaphor vs Omni

Two governed models, different scope

Omni’s product centers on a governed shared model with workbook-level extensions that promote back into it, plus a bidirectional dbt integration. Among warehouse-first BI tools, its governed-model design is well executed.

The comparison with Semaphor comes down to architecture and scope. Omni assumes a modern cloud warehouse and delivers dashboards. Semaphor connects to warehouses and operational databases alike, self-hosts when you need it to, and keeps going past the dashboard: a Data App SDK for complete governed data apps, briefings that deliver what changed on a schedule, and agents as first-class builders rather than chat assistants on top.

Pricing

Omni prices through enterprise consultative deals, sized to seats and scope in the sales cycle. Semaphor uses flat platform pricing with no per-user fees, so adding analysts, viewers, or embedded customers does not move the number, and the free tier includes embedding, so your team can validate the integration hands-on before any contract conversation.

Embedding

Both platforms embed dashboards the same way, and it is the right way: a signed URL in an iframe, with SSO, theming, and JavaScript event hooks. Omni executes this well. So does Semaphor: sign a JWT server-side, drop in the URL, and the dashboard renders fully white-labeled in any framework.

The difference is what you can ship past the embedded dashboard. Semaphor’s Data App SDK composes governed queries, filters, and visuals into complete data applications in your own React code, and your coding agent can plan and build those data apps over MCP. Custom visuals are React components you publish through the plugin system. And briefings deliver what changed, why, and where to look next, on a schedule, so your customers hear from their analytics instead of remembering to visit it.

Semantic model

Omni’s two-layer model is a thoughtful design: a governed shared model, per-workbook extensions for one-off logic, and promotion of workbook logic into the shared model once it proves reusable. The dbt integration is bidirectional; you can push BI-layer logic back into your dbt repo.

Semaphor’s Semantic Domains solve the same governance problem with less ceremony: datasets, relationships, calculated metrics and dimensions defined once, with auto-join resolution (including multi-hop) so end users and agents compose queries without writing SQL. Setup is AI-assisted, and the same domain serves dashboards, the explorer, MCP tools, and briefings.

Multi-tenancy and security

Omni implements row-level security with access filters: user-attribute values injected into the WHERE clause of every query, passed through the embed URL or POST body, with connection roles controlling data access per user. It is a proven, Looker-style pattern.

Semaphor covers the same row-level pattern and adds two coarser-grained isolation levels that SaaS vendors often end up needing: schema-per-tenant and connection-per-tenant. All three are enforced at query build time on every query path, including agent-generated queries, on all paid tiers.

Semaphor also ships the management model around the filters. Organizations, tenants, and tenant users are first-class objects: each tenant carries its own users, roles, groups, sharing boundaries, and defaults, down to per-tenant fiscal calendars and number formats, and organization-level appearance controls white-label the whole surface. Tenant users build and share dashboards strictly inside their tenant, enforced server-side. In Omni, the tenant exists as user attributes and connection roles on embed users; the administrative shell around them is yours to build.

Version control

Omni has real git integration for the shared model: repo sync, branches, and a PR-required mode. Dashboards are further behind; document APIs and content import-export endpoints for migration are still in beta, so dashboards-as-code is partial today.

Semaphor version-controls the dashboards themselves. Connect GitHub once, commit from the toolbar, review diffs in GitHub, restore any version, and promote dashboards across dev, staging, and production projects via the API.

AI and agents

Omni’s Blobby assistant answers questions and builds workbook queries and dashboards, agentic workflows plan multi-step analyses, and an MCP server exposes querying to external tools, grounded in the semantic model.

Semaphor shares the same conviction that AI must be grounded in a governed model, and goes a step further on the builder side: agents do not just query, they build. Claude or Codex can create dashboards and full data apps through MCP, with every number verified against a live governed query, and the results land in GitHub like any other code change. Briefings then keep humans informed proactively.