Find My Model

• 

  Dashboard

Choosing an AI model should not mean defaulting to the most expensive frontier model. For many products, the right answer depends on context length, image or audio support, tool calling, latency, data retrieval, deployment constraints, and cost. Find My Model turns that decision into a research-backed workflow.

The app takes a real workload or product idea, profiles the prompts and hard requirements, searches current model/provider evidence, ranks candidates from live catalogs, and then creates a recommendation with citations. Prometheux is used as the system of record for the reasoning context: one populated ontology captures requirements, candidate models, RAG settings, and evidence; a separate populated lineage concept shows how the input and research flow into the final recommendation.

The result is a practical model-selection assistant for builders who want the best fit, not just the biggest model. It helps answer: when do I need a premium frontier model, when is a faster/cheaper model enough, and what evidence supports that choice?

Built With

• clickhouse-cloud
• clickhouse-sql-http-api
• gemini
• github
• google-adk
• hugging-face-api
• langfuse
• lucide-react
• models.dev
• next.js-16
• prometheux-rest-api
• python-3.14.5
• react-19
• tavily-search-api
• typescript
• uv
• vadalog