# Dex > Let your agent build your analytics. The agent-native analytics engine for Claude Code and any agent: explore your data, build dbt models and semantic layers, and keep them in sync as schemas change. Dex is free and open source under Apache 2.0, made by Exmergo, Inc. It installs into Claude Code, OpenAI Codex, Cursor, Gemini CLI, and any agent that reads AGENTS.md, and gives the agent three skills: /dex:explore (inventory, profiling, PII detection, verified joins), /dex:transform (dbt models, tests, and semantic layers as reviewable diffs), and /dex:maintain (schema, volume, grain, and semantic drift). Dex Enterprise is the paid harness that runs Dex for a whole data team on the customer's infrastructure. ## Install - Claude Code, one command at a time: `/plugin marketplace add exmergo/exmergo-agent-plugins`, then `/plugin install dex@exmergo` - Any other agent: `npx skills add exmergo/dex` - Source: https://github.com/exmergo/dex ## Product - [Dex](https://trydex.ai/): Let your agent build your analytics. The agent-native analytics engine for Claude Code and any agent: explore your data, build dbt models and semantic layers, and keep them in sync as schemas change. - [Dex Enterprise](https://trydex.ai/enterprise): Dex Enterprise is a harness that runs open-source Dex for your whole data team, deployed on your infrastructure: Scope turns goals into scoping reports, Build turns tickets into draft pull requests, and Fix checks for drift on a schedule. Deployed on your infrastructure, never merges on its own. - [Pricing](https://trydex.ai/pricing): Dex is free and open source under Apache 2.0. Dex Enterprise brings Scope, Build, and Fix to your whole data team, priced by conversation with our team. - [Connectors](https://trydex.ai/connectors): The warehouses and semantic layers Dex connects to: BigQuery, Snowflake, Databricks, Redshift, ClickHouse, PostgreSQL, Trino, DuckDB, the dbt Semantic Layer, Apache Ossie, SLayer, Cube, Snowflake Semantic Views, and Databricks Metric Views. Every warehouse has a cost guard in its own billing unit. ## Documentation - [Dex documentation](https://docs.exmergo.com/dex): Install, connect a warehouse, the guardrails (cost controls and the PII policy), every connector, semantic layers, and the full CLI reference. The docs site has its own index for agents at https://docs.exmergo.com/llms.txt. - [AGENTS.md](https://github.com/exmergo/dex/blob/main/AGENTS.md): How an agent should use Dex, in the repository. ## Benchmarks - [Benchmarks](https://trydex.ai/benchmarks): How Dex performs on public analytics-engineering benchmarks. Every run is published under its own dated page, including the baseline it was compared against, so the numbers can be checked rather than taken on trust. - [ADE-bench](https://trydex.ai/benchmarks/ade-bench): dbt Labs' 75-task analytics-engineering benchmark, scored on whether the dbt project's tests actually pass. ([md](https://trydex.ai/benchmarks/ade-bench.md)) - [ADE-bench, 5 November 2026](https://trydex.ai/benchmarks/ade-bench/2026-11-05): With the same Claude Sonnet 5.5 model, Dex solves 60 of 75 ADE-bench tasks (80.0%) against 55 without it (73.3%): 6.7 percentage points higher, and 25% fewer failed tasks. ([md](https://trydex.ai/benchmarks/ade-bench/2026-11-05.md)) - [ADE-bench, 21 July 2026](https://trydex.ai/benchmarks/ade-bench/2026-07-21): On dbt Labs' 75-task analytics-engineering benchmark, Dex + Claude Sonnet 5 resolves more tasks than any other run we measured, and beats dbt's own published agent skills. ([md](https://trydex.ai/benchmarks/ade-bench/2026-07-21.md)) - [data-eng-bench](https://trydex.ai/benchmarks/data-eng-bench): Snowflake's 103-task data-engineering benchmark, scored pass/fail by whether the dbt project's own pytest suite runs clean. ([md](https://trydex.ai/benchmarks/data-eng-bench.md)) - [data-eng-bench, 11 August 2026](https://trydex.ai/benchmarks/data-eng-bench/2026-08-11): Dex + Claude Sonnet 5 resolves 59 of 103 tasks (57.3%) on Snowflake's data-eng-bench at k=1, the highest published figure for any Sonnet 5 configuration, and passes 94.7% of the verifiers' 2,850 individual assertions. ([md](https://trydex.ai/benchmarks/data-eng-bench/2026-08-11.md)) ## Connectors - [BigQuery](https://trydex.ai/connectors/bigquery): Google BigQuery data warehouse. Dex profiles your tables and authors dbt models directly against your BigQuery project. - [Snowflake](https://trydex.ai/connectors/snowflake): Snowflake data cloud. Dex explores schemas and authors dbt transformations with warehouse-aware cost guards. - [Databricks](https://trydex.ai/connectors/databricks): Databricks lakehouse. Dex profiles Unity Catalog tables and authors dbt models with DBU-aware cost guards. - [Redshift](https://trydex.ai/connectors/redshift): AWS Redshift data warehouse. Dex profiles your schemas and authors dbt models directly. - [ClickHouse](https://trydex.ai/connectors/clickhouse): ClickHouse Cloud and self-hosted ClickHouse. Dex profiles your tables and authors dbt models, with every scan capped by max_execution_time and max_bytes_to_read. - [PostgreSQL](https://trydex.ai/connectors/postgresql): Connect to PostgreSQL for advanced analytics, geospatial data, and complex queries on structured data. - [Trino](https://trydex.ai/connectors/trino): Self-hosted Trino, Starburst Enterprise, and Trino on Amazon EMR. Dex explores your catalogs, authors dbt models, and detects drift with cluster-load guards. - [DuckDB](https://trydex.ai/connectors/duckdb): An in-process analytical database. Point Dex at a local DuckDB file with zero credentials required. - [dbt Semantic Layer](https://trydex.ai/connectors/dbt): The dbt semantic layer (MetricFlow), local or through dbt Cloud. Dex authors and maintains your entities, dimensions, measures, and metrics as reviewable diffs, validated against a dev target. - [Apache Ossie](https://trydex.ai/connectors/ossie): Apache Ossie (incubating), the vendor-neutral standard for semantic metadata. Dex reads your .ossie.yaml documents natively and authors them as reviewable diffs, with or without a dbt project. - [SLayer](https://trydex.ai/connectors/slayer): SLayer by Motley, a semantic layer kept as YAML model files. Dex reads your SLayer models natively, authors them as reviewable diffs, and runs the SQL SLayer renders under its own cost guard, with no dbt project needed. - [Cube](https://trydex.ai/connectors/cube): Cube, the semantic layer kept as YAML model files. dex reads your cubes and views natively, authors them as reviewable diffs, and runs the SQL Cube renders under its own cost guard, with no dbt project needed. - [Snowflake Semantic Views](https://trydex.ai/connectors/snowflake-semantic-views): Snowflake Semantic Views, kept as YAML files in your repository. dex reads them with no Snowflake connection, authors them as reviewable diffs, and holds them to the warehouse when columns move. Nothing is deployed for you. - [Databricks Metric Views](https://trydex.ai/connectors/databricks-metric-views): Databricks Unity Catalog Metric Views, kept as YAML files in your repository. dex reads them with no workspace connection, authors them as reviewable diffs, and holds them to the warehouse when columns move. Nothing is deployed for you. ## Company - [Book a demo](https://trydex.ai/book-demo): Book a live demo with the Exmergo team to see Dex and Dex Enterprise on your data and use case. - [Partners](https://trydex.ai/partners): Partner with Exmergo on Dex. Referral, agency, and value-added reseller arrangements spanning Dex implementation services and Dex Enterprise. - [Blog](https://trydex.ai/blog): Insights on AI agents applied to analytics engineering, Dex product updates, and industry trends from the Exmergo team. - [Open Research](https://trydex.ai/open-research): Open-source research and tools from the Exmergo team. Explore our public repositories on AI agents for analytics engineering. - [Discord](https://trydex.ai/discord): Join the Exmergo community on Discord. Ask questions about Dex, share what you are building, and talk analytics engineering with the team that makes it. - [Exmergo](https://www.exmergo.com): The company that makes Dex. Its site also covers Exmergo Viz, which turns a governed semantic layer into dashboards. ## Optional - [Full text](https://trydex.ai/llms-full.txt): This index plus every FAQ on the site and the full record of every benchmark run, in one file. - [Privacy Policy](https://trydex.ai/privacy): How Exmergo handles personal data on this site and in Dex. - [Terms of Service](https://trydex.ai/terms): The terms for using this site and Dex.