The governed
context layer for
data teams.

ktx Cloud is the hosted, governed context layer for AI data agents. It auto-builds shared context from your warehouse, BI tools, semantic definitions, and company knowledge, then exposes that context to MCP-compatible agents and custom internal agents.

A context layer is what sits between your raw data stack and your data agents or analytics agents. It combines a governed semantic layer (metrics, entities, relationships) with schema, lineage, dashboard logic, and domain knowledge into one layer that agents can query. Without a context layer, agents hallucinate or return inconsistent answers. ktx Cloud auto-generates this context layer from your existing data stack, so your data team governs it and any agent can consume it.

ktx Cloud is built for data teams that want business users, analytics engineers, and AI agents to use the same governed metric definitions. It is especially useful when teams already have dashboards, dbt models, warehouse data, and documentation, but agents still lack reliable business context.

ktx is the open-source context layer you can run yourself. It is Apache 2.0 licensed, has no usage limits, and writes context as YAML and Markdown files in your repo. ktx Cloud is the hosted, multi-user, governed version. Same engine, plus hosted runtime, review and approval workflows, SSO, continuous ingest, observability, and included LLM credits.

Raw warehouse MCP access gives an agent tables and columns. ktx Cloud gives agents governed metric definitions, source evidence, lineage, dashboard logic, domain knowledge, permissions, and a join graph built for reliable analytics queries. Agents answer from approved context instead of guessing from schema names.

Traditional semantic layers like Cube or dbt MetricFlow require data teams to manually define every metric, entity, and relationship. ktx Cloud goes further: it auto-generates a context layer that includes a built-in semantic layer plus schema, lineage, dashboard logic, and domain knowledge. If you already have a semantic layer, ktx Cloud ingests it and fills gaps. If you don't, ktx Cloud builds one. Either way, your data team reviews and governs the result, not builds it from scratch.

Data catalogs like Atlan and Alation help data teams organize and find metadata. They don't build a context layer or power data agents. ktx Cloud auto-generates a governed context layer from your warehouse, BI tools, and domain knowledge, combining semantic definitions with schema, lineage, and dashboard logic. Agents query this context layer to deliver trusted, sourced answers.

No. ktx Cloud consumes dbt instead of replacing it. If you use dbt, ktx Cloud ingests your models, metric definitions, and documentation into its auto-generated context layer. This makes your dbt investment more valuable by exposing governed metrics to data agents that business teams can query directly. If you don't use dbt, ktx Cloud builds the context layer directly from your warehouse schemas and BI tools.

ktx Cloud runs the ktx Engine in a hosted, multi-user workspace. It reads your warehouse schemas, BI dashboard logic, metric definitions, lineage, and domain knowledge, then generates structured context: metrics, entities, relationships, and wiki pages. It continuously refines that context by learning from how teams query the data, suggesting new metrics and knowledge to fill gaps. Your data team reviews and governs the output.

ktx Cloud auto-generates metric definitions from your warehouse, dbt models, and BI tools, then surfaces them for your data team to review, approve, and enrich. As teams query the data, ktx Cloud identifies gaps and suggests new metrics and definitions to add. Each metric gets one governed definition with a clear source, so business teams always see consistent numbers.

ktx Cloud supports the same warehouse and context sources as ktx in a hosted workspace. Warehouses include PostgreSQL, Snowflake, BigQuery, MySQL, SQL Server, SQLite, and many more. Modeling and BI sources include dbt, MetricFlow, LookML, Looker, Metabase, and many more. Documentation sources include Notion and many more.

ktx Cloud can power custom internal agents, Claude Code, Claude Desktop, Codex, Cursor, OpenCode, generic .agents clients, and any MCP-compatible tool. Agents can query the same governed context layer through MCP, so every agent uses the same approved definitions and access rules.

Agents hallucinate when they lack business context and have to infer business logic from schema names. ktx Cloud gives agents approved context before they write or run SQL: governed metric definitions, source evidence, lineage, dashboard logic, domain knowledge, and a mapped join graph. This reduces repeated schema exploration, incorrect joins, and invented metric logic.

Yes. ktx Cloud supports role-based access controls with table, row, and column-level permissions that mirror your existing data governance. The data team controls exactly which metrics and data are available to which teams. Sensitive data can be excluded at any granularity, and all access is auditable.

Most data teams are up and running in under an hour. Connect your warehouse, point ktx Cloud at your dbt project or BI dashboards, and it starts auto-generating your context layer immediately. Your team reviews and governs the output. Business teams can start querying agents the same day.

Yes. ktx Cloud is SOC 2 and HIPAA compliant and uses 256-bit encryption at rest and in transit. Your data is never shared across customers or used to train models. We offer SSO, role-based access, and audit logs for enterprise deployments.