Watto

An unfamiliar object can send its owner through a long chain of searches. What is it? Which details matter? Which listings are actually comparable? Watto organizes that investigation around a photograph and a short description.

Ask the questions that matter

The application identifies the item and chooses follow-up questions relevant to its value. A ring may need a weight; a collectible may need an edition. The answers feed a research stage, followed by an estimated price range and a report that can be saved or printed.

Give each step a job

Code controls the sequence: identify, gather facts, research, appraise. AI models perform defined tasks inside that sequence. The application retains the description, answers, research, and result together, so the user can follow how the estimate was assembled. Missing information remains visible. The appraisal instructions call for wider ranges when facts are skipped, and the pipeline carries research failures into the final assessment instead of silently treating the research as complete. Current state: Local AI application with a browser interface, command-line workflow, and PDF export. Live research requires a configured model provider. The illustration above explains the workflow; it is not an appraisal result. Independent valuation accuracy and commercial outcomes have not been established. Explore a custom system