Start here

Misata builds company data, many connected tables, that shows exactly the numbers you set: MRR from $50,000 to $200,000, churn at 2% a month with a spike to 5% in one quarter, a refund rate, a channel mix. Each number is held to the cent or the row, and every dataset comes with its answer key: the SQL that reads each number back and the value it returns, month by month.

That gives you two things no hand-written script or generic generator gives:

• Data that tells a story you planned. An empty dashboard, a demo or a sales environment shows the trend you meant, with every key joining and every date in order.
• A right answer to check against. Whatever reads the data (a dashboard, an AI analyst, a text-to-SQL model, a generated query) can be graded against the key, and a wrong answer is named for the mistake it is.

The tools, and the job each one does#

Tool	Use it when you want to	Where
Studio	Describe a business in words, import your own tables, or draw them, and get a verified dataset to open, explore, change and export	Studio
The Story tab and checker	See every number you stated next to what the data holds, and check a query or a dashboard's figures against the answer key	On each dataset
The answer key	Take the ground truth with you: each metric's SQL and its value in every month, as JSON or CSV	Story tab, Export, API
Platform API	Have your own product ask for datasets on your users' behalf, load them into Postgres or ClickHouse, query them and check its own numbers	API keys and the guide
Backlot	Make an existing data pipeline (a dbt project or a Postgres schema) show a planned story in its own output, then share it as a live demo room	Backlot and the guide
MCP server	Let an AI assistant (Claude, Cursor, VS Code) design, build, query and grade datasets in a conversation	MCP and the guide
Free datasets	Take a ready multi-table dataset now, with nothing to describe	Datasets

Your first ten minutes#

1. Open Studio and choose New dataset. Start from the SaaS example, or write one sentence with the numbers that matter: "A B2B SaaS on three plans over the last two years. MRR grows from $50,000 to $200,000, with 2% monthly churn spiking to 5% in Q3 2025."
2. Read the plan before you build. Studio shows the tables, their sizes and each number it will hold. Change the story there if something is off; nothing has been made yet.
3. Build. Making and checking the rows takes seconds; the models that write the business's own content (names, catalog, wording) and the expert review take most of a minute or two. You can leave the page; the build carries on.
4. Open the Story tab. Each stated number sits beside what the data holds, month by month, with the SQL that reads it. Below it, the business's standard metrics (ARR, net new MRR, win rate) with their SQL.
5. Check a query. Paste any SQL into the checker on the Story tab. A right query is named as the metric it computes. A wrong one is told what it computes instead, with the one change that makes it right, proved by running it.
6. Take it with you. Export to Postgres, ClickHouse, Snowflake, BigQuery, Parquet, CSV, dbt seeds and more, with the answer key beside the tables.

Three ways in#

• In words: one or two sentences with the numbers that must hold. Best for a business Misata already knows (SaaS, e-commerce, banking and payments, health systems, schools, AML, machines and sensors, clinical trials).
• Your own tables: import SQL DDL, a dbt project, Prisma, Django, SQLAlchemy, Rails, OpenAPI, JSON Schema, BigQuery, Avro, DBML or CSV headers, then state the numbers on your own columns. Only the structure is read; no real values are copied.
• Drawn: lay the tables out on the canvas, pick each column's kind (an ID, a link to another table, money, a date, a choice with its shares) and see real sample values as you go.

What to read next#

• Getting realistic results: how to ask, what is exact and what is drawn, and how to read what comes back.
• The answer key and checker: the ground truth, the misreadings it names, and how to grade a dashboard or an AI analyst with it.
• What you can do with Misata: pick what you want to do, and see which tool does it and how.
• Limitations: what Misata does not do yet, stated plainly.
• Glossary: the words on the screens.