Parse

任意のブランドがAI検索にどのように表示されるかを確認

ドキュメント

Model Context Protocol

Parse MCP

Look up how any brand surfaces in ChatGPT and Google AI Overviews.

Open MCP endpointDownload MCPB

MCP 2025-06-18

Streamable HTTP

No auth

Read only

Connect

Add the endpoint to any MCP-capable client.

Claude logo

Claude Code

claude mcp add --transport http parse https://mcp.parse.gl/mcp

ChatGPT / OpenAI logo

Cursor / ChatGPT

https://mcp.parse.gl/mcp

Tools

Four primary tools call the Parse public index. All read-only.

parse_search

function parse_search(args: {
  query: string
  types?: Array<"brands" | "prompts" | "niches">
  limit?: number  // 1 to 10
}): SearchResult

Find brands, organic AI prompts, and market niches.

Use first when the user names a brand, category, or AI visibility question. Returns a ranked mix across the public index types.

Sample call

curl -sS https://mcp.parse.gl/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -H 'MCP-Protocol-Version: 2025-06-18' \
  --data '{"jsonrpc":"2.0","id":1,"method":"tools/call",
    "params":{"name":"parse_search",
    "arguments":{"query":"stripe","types":["brands","prompts"],"limit":2}}}'

Sample output

Parse research results for "stripe"

Brands
- Stripe — stripe.com
  Tool: parse_get_brand slug_or_id="stripe-com"
  Metric hint: rank score 4379, relevance 100%

Prompts
- Who offers merchant cash advances based on stripe sales volume?
  Tool: parse_get_prompt slug="who-offers-merchant-cash-advances-based-on-stripe..."
  Metric hint: rank score 37, relevance 12%

Next step: call parse_get_brand or parse_get_prompt for the most relevant result,
or rerun search with narrowed types.

parse_get_brand

function parse_get_brand(args: {
  slug_or_id: string
}): BrandBrief

Fetch a public marketing brief for one brand.

Parse score, strengths, weak spots, top prompts, citation sources, related brands, and next research questions.

Sample call

curl -sS https://mcp.parse.gl/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -H 'MCP-Protocol-Version: 2025-06-18' \
  --data '{"jsonrpc":"2.0","id":1,"method":"tools/call",
    "params":{"name":"parse_get_brand",
    "arguments":{"slug_or_id":"stripe-com"}}}'

Sample output

Stripe
Domain: stripe.com
Latest data: 2026-04-24

AI visibility summary
- Parse Score: 83/100
- Strength: 96/100
- Reach: 77/100
- Authority: 73/100
- Coverage: 272 prompts and 78 sources

Strengths
- Strongest score is Strength at 96/100.
- Top source: stripe.com (64 citations).

Weak spots and data gaps
- Lowest score is Authority at 73/100.

Top prompts to inspect
- "What is the best payment processor?" — visibility 80%
- "Best developer-friendly payment API?" — visibility 80%

Top citation sources
- stripe.com — 35% citation share across 35 prompts
- reddit.com — 8% citation share across 11 prompts
- docs.stripe.com — 6% citation share across 5 prompts

Related products to research next
- Link (brand:stripe-com-21)
- Stripe Atlas (brand:stripe-com-5)
- Stripe Billing (brand:stripe-com-2)

parse_get_prompt

function parse_get_prompt(args: {
  slug: string
}): PromptDetail

Fetch one organic AI-search prompt by slug.

Use when the user wants to inspect the exact AI-search question behind a search result.

Sample call

curl -sS https://mcp.parse.gl/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -H 'MCP-Protocol-Version: 2025-06-18' \
  --data '{"jsonrpc":"2.0","id":1,"method":"tools/call",
    "params":{"name":"parse_get_prompt",
    "arguments":{"slug":"who-offers-merchant-cash-advances-..."}}}'

Sample output

Public AI-search prompt
Who offers merchant cash advances based on stripe sales volume?
URL: https://parse.gl/prompts/p/who-offers-merchant-cash-advances...

How to use it
- Search this prompt's topic for brands and rankings.
- Compare which brands have public profiles before making
  visibility claims.

parse_get_stats

function parse_get_stats(): DatasetStats

Summarize Parse dataset scale and freshness.

Tracked brands, market questions, citation observations, and last update.

Sample call

curl -sS https://mcp.parse.gl/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -H 'MCP-Protocol-Version: 2025-06-18' \
  --data '{"jsonrpc":"2.0","id":1,"method":"tools/call",
    "params":{"name":"parse_get_stats","arguments":{}}}'

Sample output

Parse public dataset
- Brands tracked: 577,098
- Market questions: 3,850,741
- Citation observations: 47,254,155
- Last updated: 2026-03-27

What this means
- Brands are public company/product profiles Parse can resolve.
- Prompts are organic AI-search questions where buyers may
  discover or compare brands.
- Citations are source/domain observations from AI answers.

Try these

Once Parse MCP is connected, ask your assistant directly.

Look up a brand

Pull Snowflake's AI-visibility brief from Parse. Strengths, weak spots, top prompts, citation sources.

Compare brands

Compare Notion and Asana's Parse Scores and citation share. Which is gaining mindshare in AI search?

Research prompts

Find prompts about payment APIs, then inspect the strongest brand profiles and citation context.

Scan a category

Top five brands AI recommends in vertical SaaS HR tools. Then inspect each brand with parse_get_brand.

Support

Endpoint issues, schema questions, or directory submissions: [email protected].

Privacy details: see parse.gl/privacy.

{"@context":"https://schema.org","@type":"Organization","@id":"https://parse.gl/#organization","name":"Parse","alternateName":"Parse.gl","url":"https://parse.gl/","logo":"https://parse.gl/android-chrome-512x512.png"}
{"@context":"https://schema.org","@type":"WebSite","@id":"https://parse.gl/#website","name":"Parse","alternateName":"Parse.gl","description":"Parse indexes AI recommendations so brands know where they stand.","url":"https://parse.gl/","potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://parse.gl/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}