Warehouse exploration
Dex profiles warehouse tables, infers relationships, flags PII, and builds a draft map cached locally so an agent can explore a schema without manually sampling tables one by one.
Exmergo Dex
Exmergo Dex is an open-source, agent-native toolkit for exploring data warehouses, authoring dbt and semantic models, and maintaining them as schemas drift.
What Dex is
Exmergo Dex is an agent-native analytics engineering toolkit for working with data warehouses and dbt projects through coding agents. The site positions it as a tool for exploring a warehouse, transforming data into dbt and semantic models, and maintaining those models when schemas or business definitions drift.
Dex is designed to run through your own agent workflow rather than as a separate BI application. Exmergo says it works with Claude Code and other agents, supports warehouses such as Snowflake, Databricks, BigQuery, Redshift, DuckDB, and PostgreSQL, and keeps data read-only so exploration and maintenance happen without routing raw data through Exmergo.
Core capabilities
dbt transformation and semantic modeling
The toolkit can author dbt SQL models, schema tests, and MetricFlow YAML for semantic modeling, then validate against dev targets with cost guards before execution.
Schema-drift maintenance
Running /dex:maintain on a .dex snapshot detects schema drift, volume drift, and grain drift, then proposes concrete edits to reconcile the project.
Cost-guarded agent workflow
Dex is built to enforce spend controls: it uses dry runs, confirms estimated cost before running, and records activity in a local ledger.
Open-source agent skill
The product is distributed as an open-source toolkit under Apache 2.0 and is presented with command-based workflows for Claude Code and other agents.
Practical ways teams use Dex
Understanding an unfamiliar warehouse
Use Dex to inspect a warehouse schema, profile columns, infer relationships, and create a cached draft map before writing analytics logic.
Building or extending dbt models
Use Dex to turn a metric request into reviewable dbt SQL models, schema tests, and MetricFlow YAML, with validation against dev targets before changes are merged.
Detecting and fixing drift
Use Dex to run maintenance checks against a .dex snapshot when upstream schemas or row volumes change, then review the proposed reconciliation edits.
Working inside an agent-based development flow
Use Dex through Claude Code or another supported coding agent when you want analytics engineering tasks to happen in the same workflow as other development work.
Pros and Cons
Pros
- Combines warehouse exploration, dbt transformation, and drift maintenance in one workflow.
- Works with coding agents instead of requiring a separate analytics UI.
- Keeps raw data in the warehouse and uses read-only, cost-guarded operations.
- Supports a broad set of warehouse back ends mentioned on the product page.
- Open source under Apache 2.0, with a free single-player developer version.
Cons
- The source pages do not provide a full public docs set here, so some setup details and command coverage remain limited from the available evidence.
- Dex is shown as free and open source for single-player developer use; team or enterprise needs appear to sit in the broader Exmergo platform rather than the free toolkit alone.
FAQ
Is Dex free to use?
Dex is designed to work through coding agents such as Claude Code, and the pricing page says the single-player developer version is free forever. The home page also shows it as an open-source toolkit released under Apache 2.0.
What does Dex do?
Dex is presented as an analytics engineering toolkit for exploration, transformation, and maintenance. The home page shows the workflow as /dex:explore, /dex:transform, and /dex:maintain.
Which systems does Dex connect to?
The home page says Dex connects to warehouses including Snowflake, Databricks, BigQuery, Redshift, DuckDB, and PostgreSQL, and also connects to dbt for semantic-layer work.
How does Dex handle data access and security?
Dex is positioned for use through your own coding agent and runs with your own API key or subscription. The site says it is read-only against your data and that raw data and credentials do not enter the model context.
How is Dex installed or used?
The site says Dex can be used via Claude Code, npx skills, AGENTS.md, and other agent workflows such as Codex, Gemini CLI, and Cursor. The benchmark page also frames it as a plugin/skill for agents rather than a standalone BI app.
Quick Facts
- Category
- Developer Tool
- Primary use
- Agent-native analytics engineering for warehouses and dbt
- Deployment model
- Open-source toolkit used through coding agents
- License
- Apache 2.0
- Supported workflow
- /dex:explore, /dex:transform, /dex:maintain
- Source domain
- exmergo.com
Exmergo Dex Alternatives
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repos and verifies changes with compilers, debuggers, sanitizers, and tests.
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads in Firecracker micro-VMs with private networking and SDK, CLI, or MCP control.
hob
hob is an independent workspace for coding agents, with local control over sessions, terminals, history, routing, and follow-up work.
Manta AI
Manta AI is an autonomous web app testing tool that maps app behavior, catches regressions, and generates tests from a URL, no scripts or selectors needed.
SonOf
SonOf connects to your repo and PM tool, audits your codebase, and turns approved work into shipped tickets with senior engineering review.
Ghost
Ghost is a terminal-based AI assistant for chatting, code generation, and CLI tasks. Includes free models, supports Linux, macOS, Windows, and is open source.