dtex

dtex ("data extraction tool") is an open-source, self-hosted Python extract-load (EL) tool. It moves data from a source (an API, a database, a file drop) into a destination (a warehouse, a database, an object store) — and nothing more. Transformation is dbt's job.

The pitch in one line: a CLI-first, dbt-shaped extract-load tool — pipelines are configs, connectors are folders, no UI blackbox. The #1 principle is to keep it as simple as possible.

Install

pip install dtex                          # every baked connector, ready
pip install 'dtex[gcs,s3]'                # add gs:// / s3:// filesystem reads
pip install 'dtex[gcp-secrets]'           # add the GCP Secret Manager resolver
pip install 'dtex[aws-secrets]'           # add the AWS Secrets Manager resolver
pip install 'dtex[vault]'                 # add the HashiCorp Vault resolver

pip install dtex ships every baked source and destination — DuckDB, BigQuery, the filesystem source's local + Parquet path, the REST / Postgres / ShipHero / Stripe sources, the engine, the CLI. Extras stay opt-in for the cloud-storage paths of the filesystem source (gs:// / s3://) and for secret managers (only relevant if your profiles.yml uses secret:// URLs).

dtex requires Python 3.11+. It installs both a CLI (dtex) and an importable library (import dtex).

Usage

dtex init my_project                      # scaffold a project
cd my_project
dtex new source my_api                    # scaffold a source connector
dtex new config my_pipeline               # scaffold a pipeline config
dtex validate                             # check everything
dtex run -p my_pipeline                   # run the pipeline
dtex runs list -p my_pipeline             # show recent run history

A pipeline is one config file binding a source + a destination + a target + params. Run it with dtex run -p <config>. The library equivalent is dtex.run(config="my_pipeline") and returns a structured RunResult.

Pre-baked connectors

Each connector ships with its own README — click through for the auth checklist, scope requirements, schema, and known limitations.

Sources:

Destinations:

Engine: per-stream commit + atomic transactions (rollback on failure), state in the destination's _dtex_state table, run records in _dtex_runs, structured JSON-lines logs per run, secret redaction, schema evolution (evolve default, strict opt-in), pipeline-level parallelism with per-destination caps.

Secret managers: GCP Secret Manager, AWS Secrets Manager, HashiCorp Vault — each as an opt-in extra.

Bundled Claude skills

dtex ships three Claude skills inside the wheel: dtex-write-config, dtex-write-connector, dtex-debug. They teach Claude (or any other agent runtime that reads .claude/skills/) the post-streams-redesign config schema, the @stream decorator pattern, and the debugging playbook. Install them into a dtex project with:

dtex skills install         # copies into ./.claude/skills/dtex/

The first dtex command run inside a project that lacks installed skills prints a one-line hint (suppressed thereafter). Run dtex skills list to see which ship with this dtex version and which are installed locally.

If you're not using Claude, just ignore the hint — the skills are inert markdown until installed.

Documentation

The full design handbook lives in docs/. Start with 00 — Vision & Naming, 02 — Architecture, 06 — Project Anatomy, 12 — Configs, and 10 — Roadmap and Scope.

Security · Contributing · Code of Conduct

• Security policy — how to report a vulnerability.
• Contributing — dev setup, PR process, how to add a connector.
• Code of Conduct.
• Changelog.

License

Apache License 2.0 — see LICENSE.