Web Search Plus
Web Search Plus gives you two MCP tools. Both return sources, never answers.
You read the sources and synthesize; the server does not generate truth claims.
web_search — multi-provider search with automatic routing. Returns ranked
results with URL, title, snippet, provider, and routing metadata.
web_extract — fetches one or more URLs and returns bounded markdown (or
HTML), with optional query-ranked spans for long pages.
When to prefer these over the built-in search
- The user wants to see which sources say something, or wants a
quality/routing report.
- The query is provider-shaped: docs/API lookups, arXiv, CVE/security,
GitHub/OSS discovery, local/shopping, news with a recency window.
- The user wants a specific provider or a specific locale.
- You need the full text of a page, not a snippet. Use
web_extract.
- The user says "wsp", "web search plus", or names a provider.
For quick, one-line factual lookups the built-in search is fine. Do not run
both for the same question unless the user asks to compare.
web_search
Required: query.
Useful parameters:
provider — auto (default) or one of serper, brave, tavily,
linkup, exa, firecrawl, parallel, you, searxng, keenable,
serpbase, querit, donsetch, octen, tinyfish. Only providers with
a configured key work; leave auto unless the user asks.
count — results, default 5, max 20.
freshness — day, week, month, year for recency.
search_type — news for a news vertical.
country / language — locale overrides.
include_domains / exclude_domains — allow or deny lists.
mode — research queries up to three providers concurrently, dedupes,
and extracts the top sources. Slower and costs more; use for deep questions.
quality_report — true to include routing class, provider scores, and
authority signals. Useful when the user asks why a result ranked where it did.
Example:
{ "query": "Hermes Agent 0.21 release notes", "freshness": "month", "count": 5 }
web_extract
Required: urls (list).
Useful parameters:
provider — auto (default) or tavily, exa, linkup, parallel,
firecrawl, you, keenable, serper, donsetch.
format — markdown (default) or html.
render_js — true for JavaScript-heavy pages when the provider supports it.
spans + spans_query — return the passages most relevant to a query instead
of the whole page. Use this for long documents.
Extraction refuses private or internal targets (loopback, RFC1918, cloud
metadata hosts). That is deliberate; do not try to work around it.
Example:
{ "urls": ["https://example.com/docs/setup"], "spans": true, "spans_query": "environment variables" }
Working pattern
web_search with a precise query. Read the result list, not just the top hit.
web_extract the one or two URLs that actually answer the question.
- Cite the URLs you used. Say when sources disagree.
- If
web_search returns an error that no provider is configured, tell the
user to run uvx --from web-search-plus-mcp==4.1.1 web-search-plus-mcp status
and set at least one provider key (see Setup). Do not guess results.
Setup (first run)
The MCP server starts through uvx, so uv must
be installed. Provider keys are read from the environment; the plugin ships none.
Recommended starter set: You.com + Serper + Linkup.
export YOU_API_KEY=... # fast current and multilingual search
export SERPER_API_KEY=... # Google-style facts, news, local, shopping
export LINKUP_API_KEY=... # source-grounded results and clean extraction
Optional: TAVILY_API_KEY, EXA_API_KEY, FIRECRAWL_API_KEY, PARALLEL_API_KEY,
BRAVE_API_KEY, KEENABLE_API_KEY, SEARXNG_INSTANCE_URL.
Check what is configured:
uvx --from web-search-plus-mcp==4.1.1 web-search-plus-mcp status
Write a starter .env template:
uvx --from web-search-plus-mcp==4.1.1 web-search-plus-mcp setup --preset starter
Full provider list, routing rules, and tool reference:
https://github.com/robbyczgw-cla/web-search-plus-mcp
1---2name: web-search-plus3description: >-4 Use the Web Search Plus MCP tools (web_search, web_extract) for source-backed5 web research: current events, prices, release notes, docs and API references,6 arXiv papers, CVEs, local and shopping queries, and reading a specific page.7 Use when the user mentions "web search plus", "websearchplus", "wsp",8 asks which sources say something, wants results from a named provider9 (Serper, Exa, Tavily, Linkup, You.com, Brave, Firecrawl, Parallel, SearXNG,10 Keenable), or asks to extract, read, or fetch a URL as markdown.11metadata:12 source: https://skillmd.com/skills/robbyczgw-cla/web-search-plus13---1415# Web Search Plus1617Web Search Plus gives you two MCP tools. Both return sources, never answers.18You read the sources and synthesize; the server does not generate truth claims.1920- `web_search` — multi-provider search with automatic routing. Returns ranked21 results with URL, title, snippet, provider, and routing metadata.22- `web_extract` — fetches one or more URLs and returns bounded markdown (or23 HTML), with optional query-ranked spans for long pages.2425## When to prefer these over the built-in search2627- The user wants to see **which sources** say something, or wants a28 quality/routing report.29- The query is provider-shaped: docs/API lookups, arXiv, CVE/security,30 GitHub/OSS discovery, local/shopping, news with a recency window.31- The user wants a **specific provider** or a **specific locale**.32- You need the **full text** of a page, not a snippet. Use `web_extract`.33- The user says "wsp", "web search plus", or names a provider.3435For quick, one-line factual lookups the built-in search is fine. Do not run36both for the same question unless the user asks to compare.3738## `web_search`3940Required: `query`.4142Useful parameters:4344- `provider` — `auto` (default) or one of `serper`, `brave`, `tavily`,45 `linkup`, `exa`, `firecrawl`, `parallel`, `you`, `searxng`, `keenable`,46 `serpbase`, `querit`, `donsetch`, `octen`, `tinyfish`. Only providers with47 a configured key work; leave `auto` unless the user asks.48- `count` — results, default 5, max 20.49- `freshness` — `day`, `week`, `month`, `year` for recency.50- `search_type` — `news` for a news vertical.51- `country` / `language` — locale overrides.52- `include_domains` / `exclude_domains` — allow or deny lists.53- `mode` — `research` queries up to three providers concurrently, dedupes,54 and extracts the top sources. Slower and costs more; use for deep questions.55- `quality_report` — `true` to include routing class, provider scores, and56 authority signals. Useful when the user asks why a result ranked where it did.5758Example:5960```json61{ "query": "Hermes Agent 0.21 release notes", "freshness": "month", "count": 5 }62```6364## `web_extract`6566Required: `urls` (list).6768Useful parameters:6970- `provider` — `auto` (default) or `tavily`, `exa`, `linkup`, `parallel`,71 `firecrawl`, `you`, `keenable`, `serper`, `donsetch`.72- `format` — `markdown` (default) or `html`.73- `render_js` — `true` for JavaScript-heavy pages when the provider supports it.74- `spans` + `spans_query` — return the passages most relevant to a query instead75 of the whole page. Use this for long documents.7677Extraction refuses private or internal targets (loopback, RFC1918, cloud78metadata hosts). That is deliberate; do not try to work around it.7980Example:8182```json83{ "urls": ["https://example.com/docs/setup"], "spans": true, "spans_query": "environment variables" }84```8586## Working pattern87881. `web_search` with a precise query. Read the result list, not just the top hit.892. `web_extract` the one or two URLs that actually answer the question.903. Cite the URLs you used. Say when sources disagree.914. If `web_search` returns an error that no provider is configured, tell the92 user to run `uvx --from web-search-plus-mcp==4.1.1 web-search-plus-mcp status`93 and set at least one provider key (see Setup). Do not guess results.9495## Setup (first run)9697The MCP server starts through `uvx`, so [uv](https://docs.astral.sh/uv/) must98be installed. Provider keys are read from the environment; the plugin ships none.99100Recommended starter set: You.com + Serper + Linkup.101102```bash103export YOU_API_KEY=... # fast current and multilingual search104export SERPER_API_KEY=... # Google-style facts, news, local, shopping105export LINKUP_API_KEY=... # source-grounded results and clean extraction106```107108Optional: `TAVILY_API_KEY`, `EXA_API_KEY`, `FIRECRAWL_API_KEY`, `PARALLEL_API_KEY`,109`BRAVE_API_KEY`, `KEENABLE_API_KEY`, `SEARXNG_INSTANCE_URL`.110111Check what is configured:112113```bash114uvx --from web-search-plus-mcp==4.1.1 web-search-plus-mcp status115```116117Write a starter `.env` template:118119```bash120uvx --from web-search-plus-mcp==4.1.1 web-search-plus-mcp setup --preset starter121```122123Full provider list, routing rules, and tool reference:124https://github.com/robbyczgw-cla/web-search-plus-mcp