Tables & columns

You don't need to design a schema upfront — just tell your agent what you want to track:

Please set up a database where I can track my language learning progress.

From that single sentence, the agent figures out the shape of your data and calls create_table five times — one table per concern:

• languages — the languages you're learning, with target level and start date.
• vocabulary — words and translations, with spaced-repetition fields like ease_factor and next_review_at.
• vocabulary_reviews — every review attempt, result and response time.
• study_sessions — daily sessions with duration, type and quality rating.
• milestones — achievements like "finished A2" or "first conversation".

Once the schema is in place, you can start using it straight away. Try something like:

Teach me 10 intermediate words in Spanish and save them to my vocabulary table.

The agent will insert the rows and you can query and aggregate them any time — filter by language, group by difficulty, track review counts over time.

#Column types

• Text — strings, freeform.
• Number — integer or decimal.
• Boolean — true / false.
• Date — ISO timestamps; agents accept relative dates and they're resolved server-side.
• Enum — fixed set of choices.
• JSON — nested objects for messy structures.

With tables in place, head to Querying & aggregating to start pulling data back out, or jump straight to Dashboards to build a live view.