# AgentWeb (remote · api.agentweb.live)

Free web search, 130M+ business records, local search, weather and travel data for AI agents.

- Trust score: 74/100 (medium)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-05

## Components

- remote · `api.agentweb.live`: 74/100 (this document), [markdown](https://verifymcp.io/servers/live-agentweb-agentweb/api.md), [page](https://verifymcp.io/servers/live-agentweb-agentweb/api)

## Channel facts

- Endpoint: `https://api.agentweb.live/mcp`
- Transports: `streamable-http`
- Auth: `none`
- Version: `1.21.1`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-05.

- **Endpoint Security**: 80/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - No authorisation is required to call this server. Every tool declares its destructiveHint and none is destructive, so open access doesn't expose one.
  - HTTPS is enforced; there's no plaintext access path.
  - The HSTS (Strict-Transport-Security) header is present.
  - DNSSEC check failed: this domain isn't protected by DNSSEC.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 71/100
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 3896 tokens (~105/item across 37 items; 37 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 27/100
  - Stability observed for 8 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 94/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 83% of tool parameters carry a description.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - All 1 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation.
  - An AI judge read all 38 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 60/100
  - Spec-recency check failed: implements MCP spec 2025-06-18; the latest is 2026-07-28.

## Install

### How do I install the AgentWeb MCP server?

AgentWeb is a hosted endpoint at https://api.agentweb.live/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add --transport http live-agentweb-agentweb 'https://api.agentweb.live/mcp'
```

### Cursor

```json
{
  "mcpServers": {
    "live-agentweb-agentweb": {
      "url": "https://api.agentweb.live/mcp"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "live-agentweb-agentweb": {
      "type": "http",
      "url": "https://api.agentweb.live/mcp"
    }
  }
}
```

### Codex

```toml
[mcp_servers.live-agentweb-agentweb]
url = "https://api.agentweb.live/mcp"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "live-agentweb-agentweb": {
      "type": "remote",
      "url": "https://api.agentweb.live/mcp",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add live-agentweb-agentweb --url 'https://api.agentweb.live/mcp' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  live-agentweb-agentweb:
    url: "https://api.agentweb.live/mcp"
```

### Netclaw

```json
{
  "McpServers": {
    "live-agentweb-agentweb": {
      "Transport": "http",
      "Url": "https://api.agentweb.live/mcp"
    }
  }
}
```

### Vellum

```bash
assistant mcp add live-agentweb-agentweb -t streamable-http -u 'https://api.agentweb.live/mcp'
```

### Other

```json
{
  "mcpServers": {
    "live-agentweb-agentweb": {
      "type": "http",
      "url": "https://api.agentweb.live/mcp"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-03 (score 74, +1)

No change was recorded against any check on this day. Stability & Change Management went from 17 to 20. That category is still filling its 30-day observation window: 5 days of observed history at the previous scan, 6 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-01 (score 73, +1)

No change was recorded against any check on this day. Stability & Change Management went from 10 to 13. That category is still filling its 30-day observation window: 3 days of observed history at the previous scan, 4 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-30 (score 72, +1)

No change was recorded against any check on this day. Stability & Change Management went from 3 to 7. That category is still filling its 30-day observation window: 1 days of observed history at the previous scan, 2 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-29 (score 71, 0)

- [functional improvement] Stability: unverified → 0.03

### 2026-10-28 (score 71)

First indexed and scored.

## MCP tools (37)

### `search_businesses` (~327 tokens)

Search businesses

Search the AgentWeb global business directory. Use agentweb_health for current counts. If a business is not in the database yet, AgentWeb tries to discover it from the web in real time (~3s) and saves it. Returns markdown prose by default. Supports text search (q), category, city/country, and geographic search (lat+lng+radius_km). ONE CALL RETURNS EVERYTHING: name, phone, email, hours, website, address, geo, trust score. Use format='json' for structured JSON.

Input parameters:

- `category` (string): Business category filter (e.g., 'restaurant', 'hotel', 'pharmacy')
- `city` (string): City name for location filter
- `country` (string): Country code (ISO 3166-1 alpha-2, e.g., 'US', 'GB', 'DK')
- `cursor` (string): Pagination cursor from a previous response's meta.next_cursor
- `format` (string): Response format. 'text' (default) returns markdown prose; 'json' returns structured JSON.
- `lat` (number): Latitude for geographic search (requires lng and radius_km)
- `limit` (number): Maximum number of results (default: 10, max: 100)
- `lng` (number): Longitude for geographic search (requires lat and radius_km)
- `q` (string): Text search query, e.g. 'noma', 'thai restaurant', 'dentist near me'
- `radius_km` (number): Search radius in kilometers (used with lat/lng)

### `local_search` (~156 tokens)

Live local search

Search local businesses from external sources; availability and freshness vary. Returns name, star rating, review count, phone, today's opening hours, website, full address, category, and geo coordinates for each result. Use when you need CURRENT info (today's hours, latest ratings) or when search_businesses doesn't have the business yet. Use for location-based business queries. Results may be cached.

Input parameters:

- `limit` (number): Maximum number of results (default: 15)
- `location` (string): City, area, or 'near me': e.g. 'Copenhagen', 'Shibuya Tokyo'
- `query` (string, required): What to look for, e.g. 'coffee shops', 'dentist', 'thai restaurant'

### `get_business` (~113 tokens)

Get business

Fetch a single business by slug or UUID. Returns markdown prose by default (~60% fewer tokens than JSON). Slugs look like 'noma-copenhagen-abc12345'. Use when you already have an identifier from a previous search. Set format='json' for structured data.

Input parameters:

- `format` (string): 'text' (default) markdown prose; 'json' structured JSON.
- `id` (string, required): Business slug (e.g. 'noma-copenhagen-abc12345') or UUID. Slugs preferred.

### `agentweb_get_short` (~101 tokens)

Get business (short)

Fetch a business in the AgentWeb compact shorthand format (~320 bytes vs 3-5 KB). Single-letter keys (n=name, p=phone, e=email, w=website, h=hours, g=[lat,lng]…). ~80% fewer tokens. Use when you just need contact info. Schema: https://api.agentweb.live/v1/schema/short

Input parameters:

- `id` (string, required): The unique business ID from AgentWeb

### `agentweb_health` (~26 tokens)

AgentWeb health

Check the health status of the AgentWeb API, including total businesses and countries available.

### `agentweb_capabilities` (~44 tokens)

AgentWeb capabilities

Get the machine-readable description of the entire AgentWeb API: every endpoint, tool, data source, and current live counts. Use to discover what AgentWeb can do.

### `contribute_business` (~248 tokens)

Contribute business

Contribute a new business to AgentWeb, OR enrich an existing one with missing data (phone, email, website, hours, address). Deduplicates automatically: same name within 100m, or a matching phone, merges into the existing record. Use whenever you find structured business data AgentWeb doesn't have yet.

Input parameters:

- `address` (object): Address fields: { street, city, postcode, country, full }
- `category` (string): Business category, e.g. 'restaurant', 'dentist', 'hotel'
- `country_code` (string): ISO 3166-1 alpha-2 country code, e.g. 'US', 'DE'
- `email` (string): Business email address
- `hours` (object): Opening hours (e.g. { monday: '09:00-17:00', ... })
- `lat` (number): Latitude (recommended for accurate dedup)
- `lng` (number): Longitude (recommended for accurate dedup)
- `name` (string, required): Business name (required, min 2 chars)
- `phone` (string): Phone number (any format; will be cleaned)
- `website` (string): Business website URL

### `report_business` (~104 tokens)

Report business

Report a business as closed, having wrong info, or being spam. Three or more 'closed' reports automatically lower its confidence score. Use when you discover a business is permanently closed, has wrong phone/address/hours, is spam, or is a duplicate.

Input parameters:

- `business_id` (string, required): The unique business ID from AgentWeb (required)
- `details` (string): Optional free-text explanation, max 500 chars
- `report_type` (string, required): Type of issue

### `agentweb_leaderboard` (~50 tokens)

Contributor leaderboard

Get the public leaderboard of top contributing AI agents (most creates, enrichments, reports).

Input parameters:

- `limit` (number): Number of top contributors (default: 10, max: 50)

### `suggest` (~118 tokens)

Autocomplete

Autocomplete / type-ahead for business names, cities, or categories. Use to disambiguate user input before searching. Returns up to 15 suggestions by relevance. Fast (<20ms).

Input parameters:

- `country` (string): Optional ISO country code to narrow suggestions
- `limit` (number): Max suggestions (default: 10, max: 15)
- `q` (string, required): Query prefix, min 2 characters (e.g. 'nom', 'copenh')
- `type` (string): What to autocomplete (default: 'business').

### `batch_get_businesses` (~87 tokens)

Get businesses (batch)

Fetch up to 50 businesses in a single call by IDs or slugs. Saves N round-trips. Supports prose (markdown), JSON, and shorthand formats.

Input parameters:

- `format` (string): 'text' (default) markdown; 'json' structured; 'short' compact shorthand.
- `ids` (string, required): Comma-separated business IDs or slugs (max 50)

### `trending_businesses` (~59 tokens)

Trending businesses

Get the most searched/popular businesses on AgentWeb right now (last 24h or all time).

Input parameters:

- `limit` (integer): Max results (default: 10)
- `period` (string): Time period (default: '24h').

### `agent_history` (~82 tokens)

My history

Retrieve this agent's interaction history: every business it searched for, viewed, or contributed to. Persists 90 days, up to 500 interactions. Recall past lookups without re-searching.

Input parameters:

- `limit` (number): Max entries (default: 50, max: 500)
- `offset` (number): Skip N entries (default: 0)

### `agent_favorites` (~46 tokens)

My favorites

Get this agent's implicit favorites: businesses it interacted with more than once, by frequency.

Input parameters:

- `limit` (number): Max favorites (default: 20, max: 100)

### `agent_profile` (~31 tokens)

My profile

Get this agent's identity and stats: API key info, tier, contribution count, request totals, remembered interactions.

### `search_flights` (~182 tokens)

Search flights

Search flights across multiple sources (booking.com + kiwi.com) and return the cheapest. Returns airlines, prices, times, stops, duration, source, and booking links. ~3s.

Input parameters:

- `adults` (number): Adult passengers (default: 2)
- `cabin` (string): Cabin class (default: ECONOMY)
- `depart` (string, required): Departure date YYYY-MM-DD
- `from` (string, required): Origin airport IATA code (e.g. 'CPH', 'JFK', 'LHR')
- `limit` (number): Max results (default: 10)
- `return` (string): Return date YYYY-MM-DD (omit for one-way)
- `sort` (string): Sort order
- `to` (string, required): Destination airport IATA code (e.g. 'FCO', 'CDG', 'NRT')

### `search_hotels` (~143 tokens)

Search hotels

Search hotels across multiple sources (booking.com + kiwi.com) and return the cheapest. Returns name, price, stars, rating, reviews, source, booking links. ~2s.

Input parameters:

- `adults` (number): Adult guests (default: 2)
- `checkin` (string, required): Check-in date YYYY-MM-DD
- `checkout` (string, required): Check-out date YYYY-MM-DD
- `destination` (string, required): City or destination (e.g. 'Rome', 'Tokyo')
- `limit` (number): Max results (default: 10)
- `max_price` (number): Maximum price per night filter
- `sort` (string): Sort by price

### `search_cars` (~213 tokens)

Search car rentals

Search real-time car rental availability and prices on booking.com via AgentWeb. Returns vehicle, supplier (Avis, Hertz, Sixt…), price, transmission, doors, mileage, pickup location, booking links. Works with any airport IATA code. ~1s.

Input parameters:

- `driver_age` (number): Driver age (default: 30)
- `dropoff` (string): Drop-off location if different from pickup
- `dropoff_date` (string, required): Drop-off date YYYY-MM-DD
- `dropoff_time` (string): Drop-off time HH:MM (default: 10:00)
- `limit` (number): Max results (default: 15)
- `location` (string, required): Pickup: airport IATA code or city name
- `pickup_date` (string, required): Pickup date YYYY-MM-DD
- `pickup_time` (string): Pickup time HH:MM (default: 10:00)
- `sort` (string): Sort by price
- `transmission` (string): Filter by transmission type

### `kiwi_search_flights` (~154 tokens)

Search flights (Kiwi)

Search flights on Kiwi.com: aggregates all airlines, often cheaper than airline sites. Returns prices, airlines, times, stops, duration, booking links. ~1.5s.

Input parameters:

- `adults` (number): Adult passengers (default: 1)
- `cabin` (string): Cabin class (default: ECONOMY)
- `depart` (string, required): Departure date YYYY-MM-DD
- `from` (string, required): Origin airport IATA code
- `limit` (number): Max results (default: 10)
- `return` (string): Return date YYYY-MM-DD (omit for one-way)
- `sort` (string): Sort (default: price)
- `to` (string, required): Destination airport IATA code

### `flixbus_search` (~128 tokens)

Search buses

Search FlixBus bus and train routes between cities in Europe. Returns prices, times, duration, transfers, booking links. ~1s.

Input parameters:

- `adults` (number): Passengers (default: 1)
- `date` (string, required): Travel date YYYY-MM-DD
- `from` (string, required): Origin city (e.g. 'Copenhagen', 'Berlin')
- `limit` (number): Max results (default: 10)
- `sort` (string): Sort (default: cheapest)
- `to` (string, required): Destination city (e.g. 'Berlin', 'Rome')

### `currency_convert` (~82 tokens)

Convert currency

Convert between currencies using live ECB exchange rates. 30+ currencies. Returns converted amount and rate. ~0.5s.

Input parameters:

- `amount` (number, required): Amount to convert
- `from` (string, required): Source currency code (e.g. 'USD', 'EUR')
- `to` (string, required): Target currency code (e.g. 'EUR', 'JPY')

### `time_now` (~128 tokens)

Current local time

Get the current date and time in a country (e.g. Denmark), supported city, or IANA timezone. Returns UTC offset, DST, Unix timestamp and timezone-data version. Fresh every call. Countries with multiple zones return choices; select one and retry. Use this for 'What time is it in Denmark right now?'.

Input parameters:

- `location` (string): Country name/code or supported city, e.g. Denmark, DK, Tokyo. Use either location or timezone.
- `timezone` (string): IANA timezone, e.g. Europe/Copenhagen. Defaults to UTC if neither field is supplied.

### `time_convert` (~135 tokens)

Convert time

Convert a local date/time between IANA zones, supported cities or countries. Rejects nonexistent clock-change times. Repeated local times require fold=0/1 or an explicit UTC offset. HH:MM uses today's date in the source zone.

Input parameters:

- `fold` (integer): First (0) or second (1) occurrence during a backward clock change.
- `from` (string, required): Source IANA timezone, supported city or country.
- `time` (string): ISO datetime, YYYY-MM-DD HH:MM, HH:MM or now (default).
- `to` (string, required): Destination IANA timezone, supported city or country.

### `weather_forecast` (~104 tokens)

Weather forecast

Get weather forecast for any city worldwide. Returns current conditions plus daily forecast up to 14 days (high/low, rain chance, UV, sunrise/sunset, packing advisories). ~1s.

Input parameters:

- `city` (string, required): City name (e.g. 'Rome', 'Tokyo', 'New York')
- `days` (number): Forecast days (1-14, default: 3)
- `units` (string): Temperature unit (default: celsius)

### `search_attractions` (~124 tokens)

Search attractions

Search real-time attractions, tours, and activities on booking.com via AgentWeb. Returns name, price, description, rating, duration, free-cancellation, discounts, booking links. ~2s.

Input parameters:

- `date` (string, required): Date YYYY-MM-DD
- `destination` (string, required): City name (e.g. 'Rome', 'Tokyo', 'Barcelona')
- `end_date` (string): End date YYYY-MM-DD (defaults to date)
- `limit` (number): Max results (default: 15)
- `sort` (string): Sort (default: popular)

### `list_recipes` (~47 tokens)

List recipes

Discover every public recipe and its parameters, authentication and cache policy. Includes news, holidays, Wikipedia, Amazon products and provider-specific travel. A listed recipe is not a guarantee of upstream availability.

### `execute_recipe` (~117 tokens)

Execute recipe

Execute a public recipe discovered with list_recipes. Pass its exact domain, recipe_name and named params. Covers all public recipes, including those without a dedicated tool. Upstream errors are returned as tool errors. Authentication follows the selected recipe.

Input parameters:

- `domain` (string, required): Catalog key, e.g. news, holidays, wikipedia.org or amazon.com
- `params` (object, required): Named parameters from list_recipes. Do not include authentication credentials.
- `recipe_name` (string, required): Recipe name, e.g. search, list or search_products

### `web_search` (~117 tokens)

Web search

General web search. mode=index searches AgentWeb's own indexed pages with a short time budget; mode=hybrid also queries external sources (the default). Returns the top web results (title, URL, snippet) for a query: general web pages, not just businesses. Use for research, finding pages, and current info. Pair with web_fetch to read a result.

Input parameters:

- `mode` (string)
- `num` (integer): Number of results (default 10, max 20)
- `q` (string, required): Search query

### `web_research` (~154 tokens)

Web research (search + read)

Web search PLUS the cleaned text of the top pages in one call (pages are fetched in parallel). Prefer this over web_search + web_fetch whenever you need to read the results, not just list them.

Input parameters:

- `entity` (string): Explicit entity for official_website candidate identification; ownership is not verified.
- `idempotency_key` (string): Reuse for a retry of exactly the same operation/input.
- `intent` (string)
- `max_chars` (integer): Characters per page (default 2500)
- `max_duration_ms` (integer)
- `max_units` (integer)
- `num` (integer): Pages to read (default 3)
- `q` (string, required): Search query

### `web_fetch` (~59 tokens)

Fetch web page

Page fetch. Fetches a public URL and returns its cleaned text content (title + body) ready to read.

Input parameters:

- `max_chars` (number): Truncate content (default 8000)
- `url` (string, required): The page URL to fetch

### `evidence_get` (~42 tokens)

Evidence get

Retrieve an exact evidence passage and its immutable capture metadata. Evidence is scoped to this application's user/workspace and retained for seven days.

Input parameters:

- `evidence_id` (string, required)

### `research_status` (~30 tokens)

Research status

Get the status, evidence and usage receipt of an accepted research request.

Input parameters:

- `request_id` (string, required)

### `research_cancel` (~40 tokens)

Research cancel

Request cancellation of an accepted research execution. In-flight usage can remain pending; retrying cancellation does not start new work.

Input parameters:

- `request_id` (string, required)

### `watch_create` (~164 tokens)

Watch create

Create an expiring, resource-limited watch for URL text changes, text appearance or an unambiguous structured product price. Baselines are silent by default. An optional webhook receives signed changes; without it, poll watch_events.

Input parameters:

- `currency`
- `expires_in_seconds` (integer)
- `idempotency_key` (string): Reuse for a retry of exactly the same operation/input.
- `interval_seconds` (integer)
- `kind` (string, required)
- `max_checks` (integer)
- `max_pages` (integer)
- `max_units` (integer)
- `notify_initial` (boolean)
- `price`
- `query`
- `terms` (array)
- `url`
- `webhook_url`

### `watch_list` (~20 tokens)

Watch list

List change watches in the current application/user/workspace scope.

### `watch_events` (~46 tokens)

Watch events

Read detected changes, supporting evidence and delivery status for a watch in the current scope.

Input parameters:

- `after` (string): next_cursor from the previous page
- `watch_id` (string, required)

### `watch_cancel` (~34 tokens)

Watch cancel

Cancel a watch and its queued callback deliveries. An already in-flight delivery may still arrive.

Input parameters:

- `watch_id` (string, required)

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/live-agentweb-agentweb/api#diagnostics

## Score history

- 2026-11-05: 74
- 2026-11-04: 74
- 2026-11-03: 74
- 2026-11-02: 73
- 2026-11-01: 73
- 2026-10-31: 72
- 2026-10-30: 72
- 2026-10-29: 71
- 2026-10-28: 71

## Common questions

### What is the AgentWeb MCP server?

AgentWeb is an MCP server listed in the public MCP registry as live.agentweb/agentweb. Free web search, 130M+ business records, local search, weather and travel data for AI agents. This page covers its hosted endpoint (https://api.agentweb.live/mcp).

### Is the AgentWeb MCP server safe to use?

AgentWeb scores 74 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the AgentWeb MCP server expose?

AgentWeb exposes 37 tools: search_businesses, local_search, get_business, agentweb_get_short, agentweb_health, and 32 more. Their descriptions and schemas cost roughly 3,855 tokens of context every time the server is loaded.

### Does the AgentWeb MCP server require authentication?

No. We connected to AgentWeb without credentials and it answered, so anything it exposes is reachable by anyone who knows the address.

### Is the AgentWeb MCP server still maintained?

AgentWeb is still listed as active in the MCP registry. We last reached this channel on 5 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

## Links

- Remote endpoint: https://api.agentweb.live/mcp
- Authorisation metadata: https://api.agentweb.live/.well-known/oauth-protected-resource/mcp
- Website: https://agentweb.live/
- Changelog RSS feed: https://verifymcp.io/servers/live-agentweb-agentweb/api.xml
- Changelog JSON feed: https://verifymcp.io/servers/live-agentweb-agentweb/api.json
- HTML version of this page: https://verifymcp.io/servers/live-agentweb-agentweb/api