Preparing Neural Aggregator
Streaming the page shell before the model grid is mounted.
Streaming the page shell before the model grid is mounted.
Public API
Use the lightweight discovery endpoint first. Reach for the heavier explorer snapshot only when you need rich merged rows, category metrics, or debug/detail fields.
/api/v1/models/indexq, provider, category, modality, source_type, only_free, max_prompt_ppm, and sort.paginate=false mode with explicit truncated and has_more signals.sort=rank (default), sort=price, sort=context, or sort=name.price_source visible so pricing provenance stays explicit.max_prompt_tokens, max_output_tokens, openrouter_best_provider, openrouter_best_endpoint, and endpoint cache freshness via openrouter_endpoints_fetched_at_utc./api/v1/models/all only when neededThe rich explorer snapshot is best for the website grid, model drawers, category-heavy payloads, and debug/detail views. It remains the heavier contract and should not be the default “one simple link” for external clients.
curl -H "Accept: application/json" "https://nnzen.com/api/v1/models/index?q=coding&modality=text&limit=20"
curl -H "Accept: application/json" "https://nnzen.com/api/v1/models/index?provider=Anthropic&paginate=false"
curl -H "Accept: application/json" "https://nnzen.com/api/v1/models/all?q=vision&include_metrics=overall,vision&view=detail&limit=50"
This quick search calls the same lightweight /api/v1/models/index route that external clients should start with.
/api/v1/health stays available as a lightweight public health check.