novada_search
Read-onlyIdempotent
Search the web using multiple engines (Google, Bing, DuckDuckGo, Yahoo, Yandex) to find relevant results including titles, URLs, and snippets. Supports options for time range, domains, and auto-extraction.
Instructions
Search the web via 5 engines (Google, Bing, DuckDuckGo, Yahoo, Yandex). Returns titles, URLs, snippets — reranked by relevance. For complex questions needing multiple sources, use novada_research instead (it's faster and more thorough).
Use for: Current events, finding URLs, fact lookup, competitive research. Set enrich_top=true to auto-extract the #1 result. Not for: Reading a known URL (novada_extract), multi-source report (novada_research). Tip: engine='duckduckgo' is 3x faster than Google and works for most queries.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | ||
| engine | Yes | Search engine to use. 'google': best general relevance (default). 'bing': good for news and local. 'duckduckgo': privacy-focused. 'yahoo': broad index. 'yandex': Russian/Eastern European content. | |
| num | Yes | ||
| country | Yes | ||
| language | Yes | ||
| time_range | No | Limit results to a time window. 'day'=last 24h, 'week'=last 7 days, 'month'=last 30 days, 'year'=last 12 months. | |
| start_date | No | ISO date YYYY-MM-DD. Return results published on or after this date. | |
| end_date | No | ISO date YYYY-MM-DD. Return results published on or before this date. | |
| include_domains | No | Only return results from these domains. E.g. ['github.com', 'arxiv.org']. Max 10. | |
| exclude_domains | No | Exclude results from these domains. E.g. ['reddit.com', 'quora.com']. Max 10. | |
| format | Yes | Output format. 'markdown': human-readable (default). 'json': structured object for programmatic agent use. | markdown |
| enrich_top | No | Auto-extract full content from the top result. Shorthand for extract_options.top_n=1. Adds ~2-4s latency. Default: false. | |
| extract_options | No | When provided, automatically extracts content from the top top_n search result URLs and appends it to each result. Eliminates a separate novada_extract call. Note: adds latency proportional to top_n * extract_latency. Use top_n=1-3 for most queries. |