"Using free open-source models locally for research with free search engines like Brave and Google" matching MCP tools:

• compare

  XFMS — Xpansion Framework Model Source

  Run a live A/B test between 2–5 user-specified models for a stated purpose. NO ranking step — the supplied model_ids ARE the candidate set. Generates 5 representative test queries from the purpose, runs them through every named model in parallel, and returns real cost, latency, and plain-English commentary on who won what. Unknown IDs are dropped with a note; if fewer than 2 IDs resolve, the call refuses. Use this whenever the user names specific models to compare (e.g. 'A/B test X and Y'). For engine-chosen candidates, use `benchmark` instead. Costs more than `rank` (10+ live LLM calls). Free-tier note: when any candidate ends in ':free', the probe is capped at 3 queries (no adaptive expansion) because free-tier rate limits often push longer probes past the deploy's 5-minute ceiling — evidence will be shallower. The commentary surfaces this when it happens.

  Connector

• proximens_oracle_search_principles

  Proximens Oracle

  Semantic search over the Proximens GEO Oracle: a curated, continuously-updated knowledge base of 3.000+ verified Generative Engine Optimization (GEO/AEO) principles, each graded by a 0-1 confidence score and traceable to a verified source. INPUT: query (natural language, 3-500 chars); optional category (one of 13 GEO categories), top_k (1-25, default 10), min_confidence (0-1, default 0.5). RETURNS: ranked principles as JSON, each with id, title, summary, category, confidence and a relevance score; Pro/Enterprise tiers additionally return full_text and source. USE WHEN you need evidence-backed answers about how AI search engines (ChatGPT, Perplexity, Gemini, Google AI Overviews, Copilot) select, rank and cite web content.

  Connector

• search_exploits

  Exploit Intelligence Platform — CVE, Vulnerability and Exploit Database

  Browse and filter exploits using STRUCTURED FILTERS ONLY (no free-text query). Use this to filter by source (github, metasploit, exploitdb, nomisec, gitlab, inthewild, vulncheck_xdb, patchapalooza, oscs, poc_monitor), language (python, ruby, etc.), LLM classification (working_poc, trojan, suspicious, scanner, stub, writeup, tool, no_code), author, min stars, code availability, CVE ID, vendor, or product. Also filter by AI analysis: attack_type (RCE, SQLi, XSS, DoS, LPE, auth_bypass, info_leak), complexity (trivial/simple/moderate/complex), reliability (reliable/unreliable/untested/theoretical), requires_auth. NOTE: To search by product name (e.g. 'OpenSSH', 'Apache'), use search_vulnerabilities instead — it has free-text query and get_vulnerability already includes exploits in the response. Examples: source='metasploit' for all Metasploit modules; attack_type='RCE' with reliability='reliable' for weaponizable RCE exploits; cve='CVE-2024-3400' for all exploits targeting a specific CVE; vendor='mitel' for all Mitel exploits.

  Connector

• flipr_game_info

  flipr-x402

  Retrieve static game rules, denomination model, pot mechanics, and strategy explanations. Free -- no payment required. Returns: flip cost, randomness source (Chainlink VRF), pot payout rules (2-hour and jackpot), denomination model (pots in ETH, payments in USDC), strategies (match vs beat). Call this first to understand the game before using other tools. [pricing: {"cost":"0","currency":"USDC","type":"free"}]

  Connector

• compare

  XFMS — Model Source

  Run a live A/B test between 2–5 user-specified models for a stated purpose. NO ranking step — the supplied model_ids ARE the candidate set. Generates 5 representative test queries from the purpose, runs them through every named model in parallel, and returns real cost, latency, and plain-English commentary on who won what. Unknown IDs are dropped with a note; if fewer than 2 IDs resolve, the call refuses. Use this whenever the user names specific models to compare (e.g. 'A/B test X and Y'). For engine-chosen candidates, use `benchmark` instead. Costs more than `rank` (10+ live LLM calls). Free-tier note: when any candidate ends in ':free', the probe is capped at 3 queries (no adaptive expansion) because free-tier rate limits often push longer probes past the deploy's 5-minute ceiling — evidence will be shallower. The commentary surfaces this when it happens.

  Connector

• search_regulations

  Velvoite — EU Financial Regulatory Compliance

  Search the regulatory corpus using keyword / trigram matching. Uses PostgreSQL trigram similarity on document titles and summaries. Returns documents ranked by relevance with summaries and classification tags. Prefer list_documents with filters (regulation, entity_type, source) first. Only use this for free-text keyword search when structured filters aren't sufficient. Args: query: Search terms (e.g. 'strong customer authentication', 'ICT risk', 'AML reporting'). per_page: Number of results (default 20, max 100).

  Connector

• search_contacts

  xmagnet

  Search for contacts by title, company, or query. ALWAYS searches saved Xmagnet contacts first (free, instant). Only calls external API if Xmagnet returns 0 results (costs credits). Examples: 'CTOs in fintech', 'John Smith at Google', 'VPs of Sales at SaaS startups'. Default limit is 50 saved results. External API fallback returns 5 per page with load_more_contacts.

  Connector

• list_clients

  Emailens MCP

  List all 15 supported email clients with IDs, names, rendering engines, dark mode support, and deprecation status. Use the returned IDs to filter other tools like preview_email or capture_screenshots.

  Connector

• search_google_xray

  LinkedIn MCP Server (Salesbot)

  Google X-Ray search for public LinkedIn profiles via Google operators (site:linkedin.com/in). Useful when you don't want to consume LinkedIn search limits. Found profiles are saved into your contacts (in a 'Google X-Ray' list, deduplicated by profile URL) and the tool returns their contact_id values. To move them into the CRM, add them to a campaign with add_contacts_to_campaign (auto-creates CRM leads) or use a CRM tool like set_deal_stage. Paginates Google results and auto-filters duplicates.

  Connector

• search_patents

  patents

  Search USPTO patent applications and grants. Use `query` for free-text keywords ("lithium battery", "crispr"). Optional structured filters: `applicant` (company name — use ALL CAPS like "APPLE INC." for best match), `filed_after` / `filed_before` (filing date range), `granted_after` / `granted_before` (grant date range). Results include title, application number, filing date, first applicant, all applicants, inventors, status, classification. Note: ODP filtering is approximate (weighted match, not strict equality) — counts and ordering are best-effort. Powered by the USPTO Open Data Portal (data.uspto.gov).

  Connector

• demo_council

  squeezeos

  Free preview of council_verdict, scoped to IWM (Russell 2000 ETF). Same JSON shape, same engines, 5-minute cache. Use this to validate output quality and integration before paying 0.10 RLUSD per call on council_verdict for any symbol.

  Connector

• lion_declare_need

  lion-x402

  FREE. The primary way agents bootstrap research on Base. Tell LION a clear, specific need in plain language (e.g. "pre-trade risk + holder concentration for Base memecoins"). LION records it, returns an immediate useful free Quick Intel sample (now with sample_enrichment_vector + tx_context_teaser for stronger value preview), and gives the exact paid upgrade path to the $0.005 flagship bundle (full numeric vectors + decoded tx receipt + calldata + volume + credits). Always include &ref= for conversion tracking from free to paid. Start free, get value, pay only when you need depth. Recommended flow: declare_need -> quick_intel sample -> paid bundle or credits.

  Connector

• get_balance

  Unphurl

  Check your pipeline check credit balance. Shows credits remaining, total purchased, total used, and lifetime free lookups count. Credits are consumed only when unknown domains run through the full analysis pipeline. Known domains (Tranco Top 100K) and cached domains (previously analysed by any Unphurl customer) are always free. If credits_remaining is 0, you can still check known and cached domains for free. To check unknown domains, purchase more credits using the "purchase" tool.

  Connector

• maps_place_details

  Google_maps

  "Hours / phone / reviews of [business]" / "Google business info for [place]" / "is [restaurant] open" — full details for a Google Place: address, phone, hours, website, ratings, user reviews. Requires a place ID from `maps_place_search`. Use after search to drill into one specific business.

  Connector

• search_catalog

  Custom Blinds Shop

  Free-text search across the full catalogue — use for open queries like 'blockout for bedroom' or 'wood venetian'. Returns id, name, category, description, and product_url. For filtering by category, colour, or dimensions use lookup_catalog instead. Pass the returned id to get_product or get_price.

  Connector

• slipstream_note

  slipstream

  Leave a note for every future agent: a gotcha, a correction to stale info, or a tip. Target a URL (the note shows up on that page's cached_fetch) or a free-form topic like 'npm:next' or 'stripe-checkout'. Write what cost you time so the next agent gets it for free. Notes are sanitized and community-moderated; spam/injection is rejected.

  Connector

• project_health

  Open Source Intelligence MCP

  Check open-source project health for a GitHub repository — stars, forks, open issues, commit frequency, last commit date, contributor count, license, and a 0-100 composite health_score (popularity + activity + maintenance + governance). The "is this project alive and worth depending on?" tool. Source: GitHub API. PAID: $0.01 USDC per query after a daily free allowance (25/day). On a 402, pay the returned Solana memo and re-call with the SAME args plus payment_tx=<signature>. agent_id scopes your allowance; an Authorization: Bearer fnet_ key bypasses it.

  Connector

• brave_web_search

  brave

  Performs web searches using the Brave Search API and returns comprehensive search results with rich metadata. To chain into local-POI enrichment, pass `result_filter=locations` and feed the resulting `locations.results[].id` values into `brave_local_search`. To chain into the AI summarizer, pass `summary=true` and feed the returned `summarizer.key` into `brave_summarizer`.

  Connector

• search_patents

  Patent Intelligence MCP

  Search U.S. patents (USPTO PatentsView) by keyword, assignee, CPC class, date range, or type. Patent search for IP and technology-landscape research, sorted newest-first, with title, abstract, assignee, and CPC codes. PAID: $0.01 USDC per query after a daily free allowance (25/day). On a 402, pay the returned Solana memo and re-call with the SAME args plus payment_tx=<signature>. Pass agent_id to scope your allowance; an Authorization: Bearer fnet_ key bypasses the paywall.

  Connector

• brave_local_search

  brave

  Brave Local Search API returns enriched information (address, phone, hours, rating) for location-search results. Access requires the Brave Search API Pro plan; currently US-only. Two-step flow: first call `brave_web_search` with `result_filter=locations` to obtain `locations.results[].id`, then pass them here. NOTE: This tool takes location IDs from a prior web-search response; if you have a free-text query, call `brave_web_search` first.

  Connector