Вход

Измеренная доля ответов 24 криптобрендов и Web3. Открытый набор данных без аудита сайта.

давно не проверялся схема сломана Удалённый · Streamable HTTP Без авторизации Веб и поиск Данные и базы ИИ и модели Официальный реестр · 9 просмотров

Что умеет

Измеренная доля ответов 24 криптобрендов и Web3. Открытый набор данных без аудита сайта.

Перевод каталога (LLM) · показать оригинал автора

Measured share of answer for 24 crypto and Web3 brands. An open dataset, not an audit of your site.

Подключение

Удалённый сервер: клиент обращается по адресу, код на вашей машине не запускается.

Claude Desktop

claude_desktop_config.json

Cursor

~/.cursor/mcp.json

Windsurf

~/.codeium/windsurf/mcp_config.json

Вставьте блок mcpServers в файл конфигурации своего клиента. Ключи не входят в файл: вместо SET_LOCALLY подставьте свои значения.

{
    "mcpServers": {
        "ai-dablock-visibility-index": {
            "url": "https://dabyte.ai/mcp"
        }
    }
}

Вызов без установки — через шлюз

Агент вызывает инструменты этого сервера через https://www.n2j.ru/mcphub/mcp: hub_call_tool {"server":"ai-dablock-visibility-index","tool":"…","arguments":{…}} Ключ не нужен, но сервер N2J ограничивает частоту анонимных вызовов.

https://dabyte.ai/mcp

Инструменты

5 схем · ≈ 959 условных токенов

get_visibility_index

The whole current release in one call: every tracked brand in SaaS & AI tools with its rank, share of answer overall and per engine, commercial intent and quadrant. Share of answer is the percentage of a fixed panel of category buyer prompts in which an engine names the brand. Use this when the question is about the field — who leads, who is absent, how the category looks. It is one response of roughly 8 KB for 20 brands, so prefer it over calling get brand visibility repeatedly. Do NOT use it for one named brand (get brand visibility is the direct answer), for movement over time (get history

Показать схему аргументов
{
    "type": "object",
    "properties": [],
    "additionalProperties": false
}

get_brand_visibility

One brand's standing in the current DABYTE release: share of answer per engine, rank, quadrant, how many panel prompts name it, and which ones. Use this when a specific brand is named. Takes a slug, not a display name — call list tracked brands first if you are unsure, or read the slug from get visibility index. An unknown slug is not a failure to hide: the error names every valid slug, so a second attempt can succeed. A brand absent from the index has not been measured at all, which is different from a measured zero. Only SaaS & AI tools brands are tracked. For the field as a whole use get vi

Показать схему аргументов
{
    "type": "object",
    "required": [
        "slug"
    ],
    "properties": {
        "slug": {
            "type": "string",
            "pattern": "^[a-z0-9-]{1,80}$",
            "description": "Brand slug, lowercase with hyphens — 'slack', 'coinbase', 'monday-com'. Not the display name."
        }
    },
    "additionalProperties": false
}

list_tracked_brands

The names and slugs of every brand in the DABYTE index — a lookup table, nothing else. No scores, no ranks. Use it for two things: to turn a brand name into the slug get brand visibility needs, and to answer whether a brand is tracked at all. Do NOT use it when you want figures — get visibility index returns the same brands with their full measurements in a single call, so calling this one first is a wasted round trip. Absence here means the brand is not measured, not that it scores zero. Covers SaaS & AI tools only; the sibling index at dablock.ai covers the other niche. Re-measured weekly, s

Показать схему аргументов
{
    "type": "object",
    "properties": [],
    "additionalProperties": false
}

get_history

Every DABYTE release ever published, as a series per brand: share of answer at each weekly measurement with the date and panel version it was taken under. Use this for any question about change — is a brand rising, when did it enter the index, how volatile is the category. Two limits decide whether an answer is honest. Figures are comparable only WITHIN a panel version: the panel is frozen between releases and a version change alters the denominator, so a difference across that boundary is not a trend. And small moves sit inside language-model noise: since panel v3 (2026-08-10) each prompt run

Показать схему аргументов
{
    "type": "object",
    "properties": [],
    "additionalProperties": false
}

get_methodology

The rules behind every figure this server returns: the exact prompt panel and its version, which engines were measured, how share of answer is scored and rounded, the resolution of the scale in percentage points, and the editorial firewall and ownership disclosure. Call this before quoting a number as evidence, before comparing two releases, or whenever a user asks how the measurement was made or who publishes it. It is the only tool that tells you how much of a difference is meaningful, which is what stops a one-step wobble being reported as a movement. It returns rules, not figures — no bran

Показать схему аргументов
{
    "type": "object",
    "properties": [],
    "additionalProperties": false
}