Novada MCP
July 21, 2026 · View on GitHub
Every tool Novada MCP exposes, grouped by category. The README shows only the handful you reach for most; this is the full list.
- 38 tools across 6 categories. Self-host (
npx novada-mcp) exposes all 38. - The hosted default surface (
mcp.novada.com) exposes 30 — the same registry minus 8 tools that don't apply to a stateless serverless endpoint (write-gated account mutations, per-process debug state, and two browser tools that need a persistent process). It's core-derived, not a hand-curated subset. - Not sure what's on your connection? Call
novada_discover— it returns this catalog live, with each tool's name, category, and status.
Search
| Tool | Best for | Returns |
|---|---|---|
novada_search | Finding pages by query, with time/domain/geo filters. Engines: google (default), duckduckgo, yandex. | Ranked results (title, url, snippet); optional auto-extracted top result |
novada_search_feedback | Rating a prior novada_search result set to improve future ranking | Thank-you/echo confirmation |
Extract
| Tool | Best for | Returns |
|---|---|---|
novada_extract | Reading one known URL, or up to 10 in parallel, including anti-bot pages | markdown/text/html/json — main content, title, links, structured fields |
Scrape
novada_scrape is the generic gateway (16 platforms, ~87 operations). The 15 novada_scrape_<platform>
tools wrap the same backend behind a closed, typed operation enum scoped to one platform, so an
agent can't call an invalid operation for the wrong platform. Every listed operation is
verified-working — operations we can't currently deliver are excluded, not left in to fail on you.
| Tool | Best for | Returns |
|---|---|---|
novada_scrape | Structured data from 16 supported platforms (~87 operations) — Amazon, Walmart, LinkedIn, TikTok, YouTube, Instagram, GitHub, X, ChatGPT, Perplexity, etc. — instead of parsing raw HTML | Clean records in markdown/json/csv/excel/html/toon |
novada_scrape_amazon | Amazon-only structured data (product, reviews, seller, bestsellers, category/brand listings) via a closed, typed operation enum — 10 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_google | Google-only raw SERP data (web search, AI Mode, Maps details/reviews, Shopping, Jobs, Hotels, Videos) via a closed, typed operation enum — 13 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_bing | Bing-only raw SERP data (web search, videos, news, shopping) via a closed, typed operation enum — 4 verified-working operations; Bing is not a selectable novada_search engine | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_duckduckgo | DuckDuckGo-only raw SERP data (web search) via a closed, typed operation enum — 1 verified-working operation | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_yandex | Yandex-only raw SERP data (web search) via a closed, typed operation enum — 1 verified-working operation | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_youtube | YouTube-only structured data (video/channel info, transcripts, comments, video/audio downloads) via a closed, typed operation enum — 13 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_instagram | Instagram-only structured data (profiles, posts, reels, comments) via a closed, typed operation enum — 7 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_facebook | Facebook-only structured data (profiles, posts, comments, events) via a closed, typed operation enum — 6 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_tiktok | TikTok-only structured data (profiles, posts, hashtag/discover listings) via a closed, typed operation enum — 5 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_x | X (Twitter)-only structured data (post details, profile lookup by username or URL) via a closed, typed operation enum — 3 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_walmart | Walmart-only structured data (product details by keyword/category URL/SKU/zip code/URL) via a closed, typed operation enum — 5 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_shein | SHEIN-only structured product data (by product ID or product URL) via a closed, typed operation enum — 2 verified-working operations (3 known backend_broken ops excluded) | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_linkedin | LinkedIn-only structured data (job listings by filters/search URL/job URL, company info by URL) via a closed, typed operation enum — 4 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_github | GitHub-only structured repository data (by repository URL or a search-results URL) via a closed, typed operation enum — 3 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
novada_scrape_perplexity | Perplexity AI's own generated answer for a query (by URL or search term) via a closed, typed operation enum — 2 verified-working operations | Clean records in markdown/json/csv/excel/html/toon (same rendering as novada_scrape) |
Crawl / Map / Site copy
| Tool | Best for | Returns |
|---|---|---|
novada_crawl | Pulling content from a bounded set of related pages (≤20) on one site | Page bodies inline (markdown/json) |
novada_map | Discovering what URLs exist on a site before deciding what to fetch | Up to 100 URLs, no content |
novada_site_copy | Mirroring an entire docs site or knowledge base to disk | Local .md files + a compact manifest (no bodies inline) |
Proxy
| Tool | Best for | Returns |
|---|---|---|
novada_proxy | Routing your own HTTP client through a specific IP type, country, city, or sticky session | Proxy URL, shell exports, or curl flag |
Browser
| Tool | Best for | Returns |
|---|---|---|
novada_browser | Full interaction — navigate, click, type, screenshot, evaluate JS — up to 20 actions per call | Action results, screenshots, snapshots; session persists via session_id |
novada_browser_flow | Simpler click/scroll/wait/type/screenshot sequences | Action results per step |
Research
| Tool | Best for | Returns |
|---|---|---|
novada_research | One call → parallel multi-source search, dedup, and extraction for a complex question | Numbered, cited source passages (extractive, not a generated summary) |
Monitor
| Tool | Best for | Returns |
|---|---|---|
novada_monitor | Detecting changes on a specific page over time | changed/unchanged + optional field-level diffs (baseline is session-scoped) |
novada_ai_monitor | Checking brand mentions on indexed AI-company domains (openai.com, perplexity.ai, anthropic.com, ...) | Per-domain sentiment, key claims, competitor mentions, source URLs |
Verify
| Tool | Best for | Returns |
|---|---|---|
novada_verify | Fact-checking a claim before citing it | Verdict (supported/unsupported/contested/insufficient_data) + confidence 0–100 |
Account
| Tool | Best for | Returns |
|---|---|---|
novada_account | Balance, plan quotas, usage history, traffic, capture logs in one call | Unified dashboard, or one section= slice |
novada_proxy_account_create | Provisioning a proxy sub-account | Created account details — WRITE, requires confirm:true |
novada_proxy_account_list | Auditing existing proxy sub-accounts | Paginated list |
novada_ip_whitelist | Managing the proxy IP whitelist (Residential, Unlimited, Static ISP) | Whitelist entries — add/del are WRITE, require confirm:true |
novada_capture_apikey | Getting or rotating the Capture API key | Current key, or new key on reset — reset is WRITE, requires confirm:true |
novada_static_ip_mgmt | Managing static ISP IPs | IP list, or purchase/renewal confirmation — open/renew are WRITE, require confirm:true |
Discover / Setup
| Tool | Best for | Returns |
|---|---|---|
novada_discover | Finding out which tool to use — call this first if unsure | Full tool catalog: name, description, category, status |
novada_setup | Validating your key on first run; never hard-errors on a missing key | Key status, balance, onboarding guidance |
novada_session_stats | Debugging your own call pattern this session | Per-tool call counts, recent calls, uptime |