# Data Contract CLI

> Open-source command-line tool to lint, test, import, and export data contracts based on the Open Data Contract Standard, maintained by the team behind Entropy Data.

Source: https://toolsfordata.com/tools/data-contract-cli/ · Updated 2026-11-13

[Docs](https://cli.datacontract.com/) · [GitHub](https://github.com/datacontract/datacontract-cli) (1.1k stars)

Best for: data engineers who want to test data in Snowflake, Databricks, BigQuery, Kafka, and more against ODCS data contracts in CI/CD.

## Overview

- Core Strength: Open-source ODCS contract tests — schema and quality checks in CI, dbt and Python
- Test Types: Partial — SQL, DataFrame, unstructured, contracts…
- Detection: Not supported
- Monitor Coverage: 6 of 11 types — as-code
- Incident Workflow: Not supported
- Quality Analytics: Not supported
- Lineage Depth: Not supported
- Governance Overlap: Not confirmed
- FinOps & AI Observability: Not supported
- AI Agent Features: Partial — auto test gen
- Developer Surface: Partial — API, Python SDK, dbt, CI/CD…
- Deployment Model: Partial — self-hosted
- Security & Access: Not supported
- Pricing: Free — open source
- Free Option: Open source, self-hosted
- Notable Caveat: Runs on demand only; no scheduler, alerts, lineage or UI (Entropy Data adds these)

## What it's used for

Data Contract CLI is used to test data against data contracts in CI/CD: you write an Open Data Contract Standard YAML file, and one command checks the schema and quality rules on your warehouse, lakehouse or Kafka topic.

## Pricing

Data Contract CLI is free under the MIT license. Entropy Data, from the same team, sells the managed data contract and marketplace platform.

## Deployment

Data Contract CLI is a Python package you run locally, in CI or from an orchestrator. For a UI and a data product marketplace on top, Entropy Data runs as SaaS or self-hosted.

## Review

Strengths:
- Free under MIT and built on the open ODCS standard.
- One command tests a contract against 20+ platforms, from Snowflake to Kafka.
- Flags breaking changes between contract versions in CI.

Caveats:
- Command line only: no UI, catalog or approval workflow.
- No scheduling or alerts: you run it from CI or an orchestrator.
- You write and maintain the YAML contracts yourself.

Recommendation: For a UI, approvals and a data contract marketplace, consider Entropy Data, from the same team. For contracts enforced on Kafka messages, Confluent Data Contracts; for YAML checks without the contract standard, Soda Core.

## Pricing

- Open source: OSS CLI
- Source: https://github.com/datacontract/datacontract-cli (checked 2026-11-02)

## Features

### Data Testing

- SQL Table Tests: Yes — Quality rules check row count, nulls, duplicates, missing and invalid values with thresholds, and SQL rules compare any query result to a threshold. ([source](https://cli.datacontract.com/quality-rules))
- DataFrame Tests: Yes — The Python API tests Spark DataFrames registered as temporary views. ([source](https://cli.datacontract.com/testing/dataframe))
- Unstructured Data Tests: Yes — Validates records from JSON files and HTTP APIs against the contract's JSON Schema; there are no LLM checks. ([source](https://cli.datacontract.com/testing))
- Table Comparison Tests: No — No command compares aggregates between two tables; changelog and breaking compare two contract files. ([source](https://cli.datacontract.com/compare-contract-versions))
- Data Diff: No — Compares contract versions, not data rows. ([source](https://cli.datacontract.com/compare-contract-versions))
- Data Contracts: Yes — Built for data contracts: datacontract test checks a data source against an Open Data Contract Standard (ODCS) YAML file with schema, quality rules and SLAs. ([source](https://cli.datacontract.com/))
- Inspect Failed Rows: Yes — The --include-failed-samples flag returns a sample of failing rows, leaving out sensitive columns. ([source](https://cli.datacontract.com/quality-rules))

### Data Observability

- Data Profiling: No — No profiling command. ([source](https://cli.datacontract.com/commands))
- Anomaly Detection: No — Thresholds are fixed values in the contract. ([source](https://cli.datacontract.com/quality-rules/library))
- Freshness Monitors: Yes — A freshness SLA in the contract is checked on each test run — a one-off check, not a continuous monitor. ([source](https://cli.datacontract.com/quality-rules))
- Volume Monitors: Yes — A rowCount rule checks row count on each test run — a one-off check, nothing is tracked over time. ([source](https://cli.datacontract.com/quality-rules/library))
- Schema Monitors: Yes — Each test run fails when a contract field is missing or its type changed — a one-off check, not a continuous monitor. ([source](https://cli.datacontract.com/quality-rules))
- Field Monitors: No — Column statistics are not stored over time. ([source](https://cli.datacontract.com/quality-rules/library))
- Distribution Monitors: No — No distribution metrics. ([source](https://cli.datacontract.com/quality-rules/library))
- Segment Monitors: No — No group-by checks. ([source](https://cli.datacontract.com/quality-rules/library))
- Custom SQL Monitors: Yes — SQL quality rules run on each test run and compare the result to a threshold — one-off checks, not tracked over time. ([source](https://cli.datacontract.com/quality-rules))
- Partition-Based Monitors: Yes — The --filter option limits a test run to rows matching a predicate, such as the latest batch. ([source](https://cli.datacontract.com/testing))
- Window-Based Monitors: No — No time-window metrics. ([source](https://cli.datacontract.com/testing))
- UI-Based Monitors: Unknown — The edit command opens a web editor for the contract file; whether every check can be set up there is not documented. ([source](https://cli.datacontract.com/commands))
- Monitors-as-Code: Yes — Checks live in ODCS YAML files kept in Git, reviewed in pull requests and tested from the command line. ([source](https://cli.datacontract.com/))
- Pipeline Monitoring: No — No pipeline monitoring. ([source](https://cli.datacontract.com/commands))

### Incident Management

- Alert Notifications: No — No email or Slack alerts; results show in the terminal or CI. ([source](https://cli.datacontract.com/scheduling))
- Incident Triage: No — No incident workflow in the CLI. ([source](https://cli.datacontract.com/commands))
- Alert Feedback: No — No alerts to give feedback on. ([source](https://cli.datacontract.com/commands))
- Incident Resolution: No — No incident workflow in the CLI. ([source](https://cli.datacontract.com/commands))
- Ticketing Integrations: No — No Jira or ServiceNow integration. ([source](https://cli.datacontract.com/commands))
- On-Call Integrations: No — No PagerDuty or Opsgenie integration. ([source](https://cli.datacontract.com/commands))

### Data Quality Analytics

- Data Quality Dashboards: No — No dashboard; publishing results to Entropy Data adds one. ([source](https://cli.datacontract.com/commands))
- Data Health Scores: No — No score; each check passes, fails or warns. ([source](https://cli.datacontract.com/quality-rules))
- Incident Resolution Metrics: No — No incident metrics. ([source](https://cli.datacontract.com/commands))

### Data Lineage

- Table-Level Lineage: No — No lineage. ([source](https://cli.datacontract.com/commands))
- Column-Level Lineage: No — No lineage. ([source](https://cli.datacontract.com/commands))
- End-to-End Lineage: No — No lineage. ([source](https://cli.datacontract.com/commands))
- Impact Analysis: No — breaking flags incompatible contract changes, but does not look at downstream assets. ([source](https://cli.datacontract.com/compare-contract-versions))

### Data Governance

- Data Catalog: Unknown — The catalog command builds an HTML page of contract files; whether it works as a searchable catalog is not documented. ([source](https://cli.datacontract.com/commands))
- Data Catalog Integration: No — Reads schemas from catalogs on import, but does not push results to them. ([source](https://cli.datacontract.com/))
- Data Classification: No — No PII detection. ([source](https://cli.datacontract.com/quality-rules))
- Data Products: No — Results are per contract; data products are an Entropy Data feature. ([source](https://cli.datacontract.com/commands))

### FinOps

- Query Performance: No — No query performance metrics. ([source](https://cli.datacontract.com/commands))
- Cost Monitoring: No — No cost tracking. ([source](https://cli.datacontract.com/commands))

### LLMOps

- AI Observability: No — No monitoring of LLM or agent outputs. ([source](https://cli.datacontract.com/commands))

### Workflows

- Scheduling: No — No built-in scheduler; the docs say to run it from cron, GitHub Actions or Airflow. ([source](https://cli.datacontract.com/scheduling))
- CI/CD Integrations: Yes — The ci command runs tests in pipelines, writes GitHub Actions annotations and fails the build on errors. ([source](https://cli.datacontract.com/commands/ci))
- Circuit Breakers: Yes — The Python library's test() returns a run whose has_passed() result the docs use to abort the pipeline in Airflow, Dagster or Prefect. ([source](https://cli.datacontract.com/python-library))
- Webhooks: Yes — The --publish option posts test results to a URL, typically Entropy Data. ([source](https://cli.datacontract.com/commands/ci))
- REST / GraphQL API: Yes — datacontract api runs a REST API with test, export, changelog and breaking endpoints. ([source](https://cli.datacontract.com/api))
- MCP Server: No — No MCP server; agents use Agent Skills and llms.txt instead. ([source](https://cli.datacontract.com/ai-agents))
- Python SDK: Yes — The DataContract Python class runs test, lint, import, export and breaking, and returns results with has_passed(). ([source](https://cli.datacontract.com/python-library))
- dbt Integration: Yes — Imports contracts from a dbt manifest, exports dbt models and sources, and syncs contract tests into a dbt project. ([source](https://cli.datacontract.com/dbt))

### AI Agents

- AI Chat Assistant: No — No chat. ([source](https://cli.datacontract.com/commands))
- Natural Language Test Generation: No — Contracts are YAML; separate agent skills can write checks from plain English. ([source](https://cli.datacontract.com/ai-agents))
- Automated Test Generation: Yes — The import command builds a contract, with schema checks, from an existing SQL, dbt, BigQuery, Glue or Snowflake schema. ([source](https://cli.datacontract.com/dbt))
- AI Issue Triaging: No — No AI features in the CLI; separate agent skills can explain failures. ([source](https://cli.datacontract.com/ai-agents))
- AI Issue Resolution: No — No AI fixes in the CLI. ([source](https://cli.datacontract.com/ai-agents))

### Security

- SSO: No — A CLI with no user accounts. ([source](https://cli.datacontract.com/api))
- SCIM Provisioning: No — No user accounts to provision. ([source](https://cli.datacontract.com/api))
- RBAC: No — No roles; the API server uses a single shared key. ([source](https://cli.datacontract.com/api))
- Audit Log: No — No audit log. ([source](https://cli.datacontract.com/api))
- Compliance Certifications: No — An open-source CLI with no certifications. ([source](https://cli.datacontract.com/))

### Deployment

- SaaS Deployment: No — Software you run yourself; Entropy Data is the hosted platform. ([source](https://cli.datacontract.com/api))
- Hybrid Deployment: No — No vendor control plane. ([source](https://cli.datacontract.com/api))
- Self-Hosted Deployment: Yes — Installs with pip or Docker and runs locally, in CI, or as a REST API server. ([source](https://github.com/datacontract/datacontract-cli/blob/main/README.md))
- Cloud-Native App: No — A Python package, not a native app in Snowflake or Databricks. ([source](https://github.com/datacontract/datacontract-cli/blob/main/README.md))

### Open Source

- Open Source: Yes — Open source under the MIT license. ([source](https://cli.datacontract.com/))

## Integrations

18 integrations, each sourced from the vendor docs.

- Data Storage: Amazon S3, Azure Blob Storage, Azure Data Lake Storage, Google Cloud Storage
- Data Transformation: Apache Spark
- Data Warehouse: Snowflake, Google BigQuery, Databricks SQL, Amazon Redshift, DuckDB, Oracle DB, ClickHouse
- Databases: PostgreSQL, Microsoft SQL Server, MySQL
- Query Engine: Trino, Amazon Athena
- Streaming: Apache Kafka

## Alternatives

- [Soda Core](https://toolsfordata.com/tools/soda-core/): Consider Soda Core for contract checks in Soda's YAML format, with a paid cloud upgrade.
- [Entropy Data](https://toolsfordata.com/tools/entropy-data/): The paid platform from the same team: a data product marketplace where contracts are tested with the CLI.
- [dbt tests](https://toolsfordata.com/tools/dbt-tests/): Consider dbt tests if your producers are dbt models: checks run in the same CI/CD.
- [Confluent Data Contracts](https://toolsfordata.com/tools/confluent-data-contracts/): Consider Confluent Data Contracts if you're on Confluent: rules live in Schema Registry and run in client serializers.
- [OpenMetadata](https://toolsfordata.com/tools/openmetadata/): Consider OpenMetadata if you want contracts attached to catalog assets in a UI, with ODCS import and export, still free and self-hosted.
- [Great Expectations](https://toolsfordata.com/tools/great-expectations/): Consider Great Expectations for Python checks without a contract format.
