Show HN: SaturnSQL Context, telling AI tools what the database means SaturnSQL launched Context, a context layer priced at €29 per user per month on its Pro plan with Guard, that feeds shared definitions, column descriptions and synonyms to AI tools including Claude, ChatGPT, Cursor, Claude Code and Windsurf through any MCP client. SaturnSQL says the product addresses AI tools guessing column meanings — such as amounts stored in cents or choosing between orders and orders_v2 — by letting teams write plain-language context that every connected tool reads, and it imports existing table and column comments from PostgreSQL, Redshift, MySQL, SQL Server, Oracle, BigQuery and ClickHouse. SaturnSQL positions Context against semantic layers, noting that definitions guide the AI's SQL rather than guaranteeing identical compiled metrics, and that teams can export their context as Ossie or dbt YAML to avoid lock-in. The context layer for AI AI tools guess what your columns mean, and the numbers come out wrong. Context teaches Claude, ChatGPT and Cursor what your data means, so they answer the way your team would. €29/user/month · Pro plan, with Guard last order at now - interval '90 days' intent: "revenue by active customer" instructions: "Amounts are in cents. Use orders v2, not orders." definitions: { name: "active customer", sql: "last order at now - …" } tables: { name: "orders v2", columns: { name: "total", description: "Order total in cents, before VAT", synonyms: "revenue" } } Every AI tool. Same answers. Your team asks Claude, ChatGPT, Cursor, Claude Code, Windsurf or any other MCP client. Without context, each one guesses what your data means, and two people asking the same question get two numbers. With Context, every one of them reads the same context files, definitions and descriptions. - Claude - ChatGPT - Cursor - Claude Code - Windsurf - Any MCP client What an AI needs from your database The schema says what exists. It doesn’t say what amounts are in, which of two tables is current, or what your team means by an active customer. | The AI sees | It also needs to know | Where it goes | |---|---|---| | orders.total bigint | In cents, before VAT | Column description | | orders, orders v2 | Use orders v2. orders is the old system | Context files | | customers.last order at | An active customer bought in the last 90 days | Definition, with its SQL | | acct mgr id | People say account manager or owner | Synonyms | | COMMENT ON TABLE customers | Already written in the database | Read as it is | How it works 1. 1 Write down what only your team knowsWhich table is current, what the amounts are in, what counts as an active customer. A sentence each, in the app or straight from a conversation with the AI. 2. 2 Every AI tool your team uses reads itClaude, ChatGPT, Cursor and Claude Code get the context that fits each question, so nobody has to explain the data again. 3. 3 Comments in the database count tooTable and column comments in PostgreSQL, Redshift, MySQL, SQL Server, Oracle, BigQuery and ClickHouse come along. Where your team writes its own, that wins, and the database stays untouched. 1. Schemaorders v2.total bigint 2. Database commentOrder total 3. SaturnSQLOrder total in cents, before VAT 4. What Claude readstotal bigint: "Order total in cents, before VAT" Correct the AI once When you tell the AI it got a column wrong, it can save the correction for everyone. The next person, and the next AI tool, gets it right without being told. Every change has a name on it. Context layer vs semantic layer Context is not a semantic layer, and doesn’t try to be. A semantic layer compiles governed metrics; Context tells an AI what your data means so it can write the SQL itself. | | SaturnSQL Context | Semantic layer | |---|---|---| | What you build first | Nothing. Write context files as questions come up | Models, metrics and joins in YAML, LookML or views | | How AI uses it | Reads the meaning, then writes SQL | Queries governed metrics through an API | | Who maintains it | Anyone who knows the data, in plain language | Usually analytics engineers, in code | | Same number every time | Guided by definitions, not guaranteed | Yes, metrics compile to the same SQL | | Time to first answer | Minutes | Once the first models are written and deployed | If finance needs revenue to match to the cent in every tool, model it in a semantic layer. For everything else, context gets AI answers right from day one. More in what is a context layer https://saturnsql.com/blog/what-is-a-context-layer . See the best semantic layer tools https://saturnsql.com/use-cases/best-semantic-layer-tools and Atlan alternatives https://saturnsql.com/use-cases/atlan-alternatives for the heavier options. No lock-in Start from the Apache Ossie model you already have, and take everything your team writes with you as Ossie or dbt YAML on every plan. If you adopt a semantic layer later, start from here. version: 0.3.0.dev0 name: prod postgres ai context: instructions: Amounts are in cents. Use orders v2, not orders. datasets: - name: orders v2 description: One row per paid order fields: - name: total datatype: Integer description: Order total in cents, before VAT ai context: synonyms: revenue, amount AI that understands your data and only sees what it should Context and Guard come together on the Pro plan, and Guard’s rules apply to context too. - Hidden tables stay hidden, in the schema and in the context. - Masked columns stay masked in everything the AI reads. - Turn AI off for a database, and its context goes dark too. More on SaturnSQL Guard https://saturnsql.com/guard and database governance for MCP and AI https://saturnsql.com/use-cases/mcp-governance . Questions about Context - What is a context layer for AI? - A context layer is what an AI tool reads about your data besides the schema: what tables and columns mean, which units they use, which table is current, and how the business defines its terms. With it, an AI writing SQL picks the right table and the right filter instead of guessing from column names. - What is SaturnSQL Context? - SaturnSQL Context is a context layer for AI on the databases you already have. Your team writes down what the data means, as context files, definitions, descriptions and synonyms, and Claude, ChatGPT, Cursor, Claude Code, Windsurf or any other MCP client reads it through SaturnSQL’s MCP server, together with the table and column comments already in your database. It comes with the Pro plan, €29 per user per month, together with Guard. - What is the difference between a context layer and a semantic layer? - A semantic layer, such as dbt’s, Cube or LookML, defines metrics, joins and models in code, and tools query those metrics so a number comes out the same every time. It needs modelling first. A context layer describes the data in plain language so an AI can write the SQL itself. It works from day one but does not guarantee identical numbers. Many teams start with context and add a semantic layer for the few metrics that must never drift. - Do I need a semantic layer for AI? - Not to get useful answers. Most wrong AI queries come from missing meaning, such as amounts in cents or an old table that looks current, and plain-language context fixes those. If finance needs governed metrics that match to the cent across every tool, a semantic layer is the better fit, and you can export SaturnSQL Context to Ossie or dbt YAML as a starting point. - Which databases does Context work with? - Context reads table and column comments on PostgreSQL, Redshift, MySQL, SQL Server, Oracle, BigQuery and ClickHouse. Context files, definitions, descriptions and synonyms work on every connection SaturnSQL supports. Snowflake and Databricks are not covered. - Can an AI tool update the context? - Yes. Through the MCP tools set description, set descriptions up to 200 tables and columns at once , write context file, upsert definition and delete definition, an AI tool can save what you explain in a conversation, and import ossie brings in a whole Apache Ossie model after showing a preview. It needs the person to own the connection or be an admin, a plan with Context, and an API key or sign-in with the write scope. Each change is recorded with the person’s name. The same changes can be made over the REST API at /api/context, and export ossie gives an AI tool the context as Ossie YAML. - Why does an open format matter when AI changes so fast? - Because the AI tools, models and standards will keep changing, and what your data means mostly won’t. If that meaning lives in one vendor’s closed format, every switch means writing it again. Kept as Apache Ossie YAML, it moves with you to the next tool, and you can keep it in git and review it like code. - Can I export my context? - Yes, on every plan. Export a connection as Apache Ossie YAML, the open semantic model format formerly called Open Semantic Interchange, or as a dbt schema.yml with models or sources. Context files and definitions go along, so nothing you write is locked in. On the Pro plan you can also import an Apache Ossie model, and SaturnSQL shows what will change before it saves. - What does Context not do? - It is not a metrics engine: definitions guide the AI, which still writes the SQL. It does not crawl lineage or catalogue other tools, it does not describe relationships between tables yet, and it does not cover Snowflake or Databricks. If you need those, a semantic layer or a data catalog is the better choice. - How much does Context cost? - Database comments reach AI tools on every plan with the MCP server, from Starter at €19 per user per month. Writing your own context files, definitions, descriptions and synonyms comes with the Pro plan at €29 per user per month, together with Guard. Start a 14-day trial from the pricing page. Connecting an AI tool? Start with the SaturnSQL MCP server https://saturnsql.com/use-cases/mcp-server . Give your AI tools the meaning, not just the schema Explain your data once, and every AI tool your team uses works from it. €29/user/month Pro plan: everything in Starter, plus - Context: context files, definitions, and table and column descriptions every AI tool reads with the schema - AI tool access per connection: off, shared queries only, read-only or same as the user - Approvals: risky AI tool requests wait for a person, alerted in Slack or by email - Limits on AI tool queries for each database or group: rows per query, queries a minute per person, timeout, BigQuery bytes - Access rules per group: hide tables, hide or mask columns, filter rows - Audit log of every query, from people and AI - Query version history