Built on the Open Data Contract Standard

Ship data your team can trust.

Catalyst checks the data in your warehouses against contracts you write in ODCS. Run the checks on a schedule and find bad data before it shows up in a dashboard.

• No credit card
• Google & Microsoft SSO
• Read-only access

Supported platforms

PostgreSQL, MySQL, SQL Server, BigQuery, Microsoft Fabric, Amazon Redshift

What Catalyst does

Connect a warehouse, write a contract and read the results, all in one place. Contracts use ODCS, so they aren't locked into our product.

• Connect your warehouse

  BigQuery, PostgreSQL, MySQL, SQL Server, Redshift and Microsoft Fabric, plus CSV, JSON and Excel uploads. Connections are read-only and take a few minutes to set up.

• Import schema instantly

  Catalyst reads information_schema and pulls in your columns, types and keys automatically.

• Rules in ODCS

  Not-null, uniqueness, range, pattern, freshness and referential-integrity checks, stored as ODCS quality rules.

• Builder & YAML in sync

  Edit rules in the visual builder or straight in the YAML contract. A change in one shows up in the other, and the YAML is the master copy.

• Schedule & run

  Trigger validations manually or on a cron schedule, and see pass, warn and fail results the moment a run completes.

• Results & coverage

  Drill into failing rows, track pass/warn/fail history, and see coverage across every quality dimension.

Built to keep your data where it is

Catalyst reads metadata, never moves your data, and signs everyone in through your existing identity provider.

Single sign-on (Google · Microsoft)

Sign in with the identity provider your team already uses. Access follows your directory: remove someone there and they lose access to Catalyst too. There are no extra passwords to manage.

• Single sign-on

  Sign in with your Google or Microsoft account instead of yet another password.

• Read-only access

  Catalyst only needs metadata and read access. Checks run inside your warehouse and only the results come back.

• Encrypted in transit and at rest

  Encrypted in transit over TLS 1.2+ and encrypted at rest for all metadata and configuration.

• Granular RBAC

  Admin, Editor and Viewer roles scoped per workspace, managed in settings.

• Audit log

  Every rule change, contract edit and validation run is recorded and exportable.

• Credential isolation

  Warehouse secrets are encrypted with rotating keys, decrypted only to run a check, and never returned by the API.

Ask Catalyst from Claude, ChatGPT or your editor

Catalyst runs an MCP server, so the AI assistant you already use can read your contracts and check results, and on Team and Enterprise it can run checks and edit rules too. You connect it with a URL. There is nothing to install.

Things you can ask

• Which datasets failed last night, and why?
• Add a not-null rule on customers.email and re-run the check.
• Show me the current contract for orders and suggest missing rules.
• Scan contacts for duplicates and summarise the biggest clusters.

Server URL: https://ctlst.eu/mcp

Connect your client

• claude.ai and Claude Desktop: Add a custom connector, paste the URL and sign in.
• ChatGPT: Add a custom connector in developer mode and sign in.
• Claude Code: One command with a personal access token.
• Cursor and VS Code: An entry in mcp.json with a personal access token.

Limits that always apply

• Failing-row samples stay off until a workspace admin turns them on
• No access to connection credentials
• No deletes
• Custom SQL rules can be run, not created or edited
• Every change is in the audit log
• Tokens expire (90 days by default, 365 at most) and can be revoked at any time

Read access on every plan, including the free Starter plan. Running checks, editing rules and contracts, and duplicate scans need Team or Enterprise. Runs count towards your monthly run quota. Read the setup guide

Try Catalyst on sample data, no signup required

One click opens a workspace that's already set up: a sample warehouse, imported schemas, ODCS rules, schedules and past run results.

• Sample warehouse
• 6 datasets + schemas
• ODCS rules & contracts
• Scheduled run history

Simple, per-seat pricing

Start on the free plan and pay per seat as the team grows. The price doesn't depend on how much data you check.

• Starter — €0, free forever

  For individuals validating a first dataset.

  • 1 connection · 1 dataset
  • PostgreSQL & MySQL
  • Manual validation runs
  • Single user
  • AI assistants (MCP): read access

• Team — €29, per user / month

  For data teams shipping trusted pipelines.

  • 10 connections · 50 datasets
  • All warehouse connectors
  • Scheduling & reports
  • Builder + YAML contracts
  • Unlimited team members
  • AI assistants (MCP): read and write

• Enterprise — Custom, billed annually

  For platform teams at scale.

  • Unlimited everything
  • Google & Microsoft SSO
  • Audit log & RBAC
  • SLA & dedicated support
  • On-prem / VPC option
  • AI assistants (MCP): read and write

Frequently asked questions

What teams ask us before they connect their first warehouse.

• How do I validate data in my warehouse?

  Connect the warehouse read-only, let Catalyst import the schema, and put rules on the columns that matter — not-null, uniqueness, ranges, patterns, freshness, referential integrity. Catalyst compiles those rules to SQL that runs inside your warehouse, on demand or on a cron schedule. Every run keeps its pass/warn/fail history, so you can see the day a dataset started drifting instead of hearing about it from a dashboard.

• What is a data contract?

  A data contract is a machine-readable agreement about a dataset: which columns exist, what types they are, what the values are allowed to look like, and how fresh the data has to be. Because it is a file rather than a wiki page, it can be reviewed, versioned in Git and enforced automatically on every run. In Catalyst the contract file is the master copy. The rule builder and the YAML editor both edit that same file.

• What is ODCS?

  ODCS is the Open Data Contract Standard: an open YAML specification for describing a dataset's schema, ownership, service levels and quality expectations. It is developed under the Linux Foundation's Bitol project rather than by a vendor, so no single tool controls it. Catalyst reads and writes ODCS natively, which means your contracts stay portable and readable outside our product.

• Which databases does Catalyst support?

  Catalyst connects to BigQuery, PostgreSQL, MySQL, SQL Server, Redshift and Microsoft Fabric, and it validates CSV, JSON and Excel files you upload. Every connection is read-only and needs nothing beyond metadata and select rights. The free Starter plan covers PostgreSQL and MySQL; the Team plan unlocks every connector.

• Does Catalyst move or copy my data?

  No. Validation queries run inside your own warehouse, and only the results come back: pass or fail, the number of violating rows, and up to five example failing rows per check so your team can see what went wrong. Nothing else is copied out. Warehouse credentials are encrypted at rest, decrypted only to run a check, and never returned by the API.

• Can I run validations on a schedule?

  Yes. Any contract can run on a cron schedule as well as on demand, and the pass, warn and fail results are on the dashboard the moment a run completes. Each run is stored with its results, so coverage and quality trends are visible per dataset and per quality dimension.

• How is Catalyst priced?

  Starter is free forever: one connection, one dataset, manual runs. Team is €29 per user per month, or €290 per user per year, and includes every connector, scheduling and reports. Enterprise is custom-priced for platform teams that need SSO, audit logging, an SLA or an on-prem deployment. All prices exclude VAT.

• How long does it take to get started?

  Usually a few minutes. Adding a connection and importing a schema takes a few clicks, and the first rules are generated from the columns Catalyst finds. If you would rather not connect anything yet, the one-click demo opens a workspace with sample datasets, contracts and run history. You don't need an account.

• Can I use Catalyst from Claude or ChatGPT?

  Yes. Catalyst has an MCP server at https://ctlst.eu/mcp. In claude.ai, Claude Desktop and ChatGPT you add it as a custom connector and sign in with your Catalyst account. Claude Code, Cursor and VS Code connect with a personal access token. Reading datasets, contracts and run results works on every plan, including Starter. Triggering runs, editing rules and contracts and starting duplicate scans needs Team or Enterprise.

• What can an AI assistant see in my workspace?

  Only the workspace you connected it to, and only what your role allows there: dataset names, contracts, rules, run results and duplicate checks. It never sees connection credentials. Failing-row samples are off by default and are only included if a workspace admin turns them on. The assistant cannot delete anything or add custom SQL rules, and every change it makes is recorded in the audit log.

See what Catalyst finds in your data

Launch the demo in one click, or talk to us about rolling Catalyst out across your data platform.