Trusted in production by data teams, including
The problem
Context drift does not fail CI
The loop
Serve always current, tested, approved context to your agents
Keeping enterprise-scale context up to date
“In one summer, I built the context for our analytics agent on Alan's 6,900-table warehouse, on my own. Using Cassis tools in Claude, I automatically draft each domain from our existing data assets, run checks and evals, review the PR, make the decisions, and move on to the next domain.”
Detection
How Cassis detects drift
As the context runtime for your analytics agents, Cassis records a gap each time an agent has to guess: no governed definition, two plausible signals, a clarifying question.
It also diffs every schema and dbt change against the context that references it. Your definitions are imported, not redefined.
When nothing your team approved covers a question, the agent says so and asks instead of guessing.
Your workflow
Manage your context directly in your Git repository
Cassis structures your context as a Markdown and YAML file tree following Google's Open Knowledge Format.
Use the CLI and MCP from your preferred development environment to inspect detected issues, review and test proposed fixes, and publish approved context through your existing Git workflow.
The repository above is a specimen. Browse a real context repo on GitHub, CI gates included.
Where it fits
One approved version for every agent you run
The context lives in your Git repository, not inside a warehouse or one assistant. Every agent you run reads it through Cassis over MCP, so they all get the same approved version, and a gap exposed by any agent's question comes back to your team.
- 01
Connect your environment
Connect Cassis to your Git repository and install the Cassis CLI. Then give it your schema: a scoped read-only warehouse connection, or a DDL dump if you would rather not connect one yet.
- 02
Bootstrap your context
Run the bootstrap from your coding agent. Cassis writes the first context tree to your repository from your schema, dbt project, BI exports, query history, and docs.
- 03
Serve it over MCP
Connect Claude Code or any MCP-compatible agent. Cassis can answer questions, generate SQL, or return only the context the calling agent needs.
- 04
Maintain it in your workflow
List detected issues from your development environment, apply and test proposed fixes, then merge the approved pull request to publish a new context version.
Security
Choose your security posture
Data access
Metadata only
No warehouse connection
Cassis uses your schema, dbt project, documentation, and questions. Your warehouse or existing agent executes the generated SQL. Row data stays in your environment. Evals score by SQL equivalence review.
Scoped warehouse access
Selected datasets only
Cassis executes generated SQL through a dedicated read-only role. Access is limited to the datasets your agents use, with no warehouse writes. Evals execute the generated and expected SQL and compare result rows.
Hosting
Cassis-managed
Hosted by Cassis
Hosted on AWS in Paris, in the eu-west-3 region.
Cloud-prem
Available on request
A Cassis-managed private deployment in your own cloud environment.
Inference
Managed inference
Anthropic through Bedrock
Cassis manages Anthropic inference through Amazon Bedrock.
BYOK Bedrock or Vertex AI
Your own account
Use your Amazon Bedrock or Google Vertex AI account.
Connecting your own warehouse starts with a conversation. One domain's schema is enough, as a read-only connection string or a DDL dump. Talk to us.
Blog