See how users chat with the AI assistant and how they use the MCP server.
User asks the assistant to generate a realistic synthetic e-commerce database from a prompt; the assistant discovers model options, starts a background conversation, polls until complete, and returns the generated databases.
User has an existing SQLite schema and wants Fabricate to populate it with realistic test data; the assistant uploads the file, seeds a project, and runs a generation conversation against the real schema.
An account owner audits seat utilization, API-key usage, and per-user/per-key API call volume for the current month to plan license renewals and spot heavy programmatic callers.
User runs a reusable export workflow to produce a filtered CSV extract from a project's database, parameterized by date range and region, then downloads the resulting file.
User iterates on an existing dataset via follow-up messages, then decides a change went too far and asks the assistant to roll back the database to the state before that change using revert_conversation.