Sema

Sema is a semantic layer that helps teams query business data in plain English through a governed semantic graph. It suits organizations with complex warehouses and stricter security needs; its multi-seat paid plans make it a poor fit for a solo buyer.

Its strongest case is making answers inspectable and reusable, while its narrow current connector coverage and steep entry point for paid use limit its reach.

Overview

Sema connects databases, infers data meaning, and answers business questions in natural language. Tables, joins, metrics, and business terms form a typed graph, a useful foundation for teams trying to keep definitions consistent across warehouses with thousands of tables. Verified inferred joins can be reused in later answers, reducing the need to establish the same relationships repeatedly.

Answers can include an automatically selected chart, and Sema can ask clarifying questions before proceeding. Showing the SQL and graph nodes behind an answer gives analysts a way to review its path instead of relying on an unexplained result. That transparency is valuable, but it does not remove the need for people to check whether the underlying data and definitions are appropriate.

Key features

Beyond conversational queries, Sema supports governed metrics, dimensions and joins, row-level security, and a query API. Users can save reports, schedule email delivery, and configure metric alerts for thresholds or deviations. These functions make it more than a chat interface, though the free Sandbox excludes scheduled delivery, alerts, and signals.

Connector coverage is currently strongest for SQL databases: PostgreSQL, MySQL, MariaDB, SQLite, and any SQLAlchemy URI are available. ClickHouse and Amazon Redshift are in beta; SQL Server, Snowflake, BigQuery, Databricks, and dbt metadata are on the roadmap. Teams dependent on those roadmap systems should not treat Sema as a fit until the needed connection is available.

Sema exposes its capabilities as MCP tools for Claude, Cursor, and in-house agents. It supports OAuth2/OIDC identity providers including Okta, Entra, Auth0, and Google Workspace. Security features include TLS in transit, encryption at rest, role-based access, restricted-column masking, workspace isolation, and an append-only log of queries, plans, and configuration changes. SOC 2 Type II is in progress; GDPR-ready data processing terms are available, and HIPAA support with a BAA is available for eligible plans on request.

Deployment can run in the customer’s VPC or Sema’s environment. The self-hosted Kubernetes option uses a Helm chart and customer-provided Postgres and object storage, giving infrastructure control to organizations prepared to operate those dependencies.

Pricing

Sandbox is free, billed Free forever. It allows one connector, one user, 100 chat queries per month, and five saved reports; embeds carry “Powered by Sema.” There is no scheduled delivery, alerts, or signals. Those caps make it useful for a limited evaluation, not sustained team reporting.

Pro is $49 a month, billed $49/user/mo, with a three-seat minimum and a 14-day trial. It adds unlimited connectors, chat, reports, schedules, alerts, and signals, plus RBAC, an audit log, and email support. The minimum means the practical starting commitment is three paid seats, even for a small team.

Business is $99 a month, billed $99/user/mo, with a ten-seat minimum. It includes Pro, SSO, advanced governance (Compliance Agent — V2), AI eval harness and federated joins (V2), priority support, and a 99.9% SLA. The seat floor makes it aimed at larger teams; features marked V2 are not presented as current capabilities.

Enterprise has custom pricing, billed annually. It includes Business, SAML/SCIM, custom DPA, BYOK encryption, an on-prem option using Helm, SOC 2, HIPAA, and GDPR evidence packs, a named customer success manager, and a 99.95% SLA. It is the plan for organizations that need those deployment, identity, or assurance provisions.

Platforms

Sema supports API, self-hosted, and web use. The hybrid deployment model offers a choice between running in the customer’s VPC and Sema’s environment, while Kubernetes self-hosting requires customer-provided Postgres and object storage.

Who it's for

Sema is best suited to enterprise and regulated teams that need governed answers across large data warehouses, inspectable query paths, and scheduled reporting. Teams using its available SQL connectors can make the most of that combination. It is a weaker choice for solo users, small teams unwilling to meet seat minimums, or organizations whose required warehouse connectors remain in beta or on the roadmap.

Pros and cons

• Pros: SQL and traversed graph nodes are shown with answers, making query behavior more reviewable.
• Pros: A typed graph and reusable verified joins support consistent metrics and relationships across complex warehouses.
• Pros: Reporting, scheduled alerts, row-level security, audit logging, and VPC or self-hosted deployment address needs beyond conversational querying.
• Cons: Sandbox is limited to one user, one connector, and 100 monthly queries, and lacks scheduled delivery and alerts.
• Cons: Pro requires three seats and Business ten, putting paid access out of reach for many solo buyers and small teams.
• Cons: Several major warehouse connectors, including Snowflake and BigQuery, are still on the roadmap.

Alternatives

For a broader comparison, browse Semantic Layer Software. Choose Cube if its free plan’s five workbooks, 1,000 daily requests, and one-day query history suit your workload. Starburst Enterprise Context Layer is another option, with a free plan for up to three clusters and ad hoc queries in standard execution mode.

dbt Semantic Layer may fit a developer who can work within its free plan of one seat, 3,000 successful models per month, and one project, which does not include the Semantic Layer. Strata offers a free plan with one developer, 25 users, and two data sources per project, making it worth considering for a project with those limits.

Definite may suit a small team that can work within its free plan of two users, two connectors, 1 GB storage, five monthly credits, and daily sync. Malloy is a free alternative. Verdonz has a free plan limited to a guided workspace with sample data and no own business data. MetricFlow is a free, Apache 2.0-licensed option for teams with a working dbt project and dbt adapter.

Verdict

Choose Sema if your organization needs governed, inspectable natural-language queries across complex data and can use its current connectors. Its graph, reporting, and deployment options form a credible enterprise-oriented package. Look elsewhere if you need a low-cost solo plan, broader connector availability today, or cannot justify the three-seat Pro minimum.