MCP

Brandsearch ships an MCP server so any AI tool that speaks the protocol — Claude Desktop, Cursor, ChatGPT (custom GPTs), Windsurf, Copilot, and more — can search brands, ads, products, and analytics directly from the chat.

Why MCP instead of pasting an API URL? Both work. The llms-full.txt approach gives the AI documentation it can use to call the API itself. MCP gives the AI structured tools with typed arguments — the AI doesn't have to write code, it just calls search_brands(niche="Beauty & Skincare") and gets results back. Less rope, fewer mistakes.

Endpoint

https://api.brandsearch.co/v1/mcp

Authentication is via OAuth. Sign in once from the AI tool and it stays connected for 30 days — no manual API key management on the client side.

Connect from popular AI tools

Every tool uses the same MCP URL — https://api.brandsearch.co/v1/mcp — and the same OAuth sign-in. Copy it once and follow the steps for your tool below.

Claude

Open Claude connectors

1Click the button below to open Claude's connector settings.
2Click 'Add custom connector' and paste the MCP server URL above.
3Authorize Brandsearch when Claude prompts you.

ChatGPT

1Open chatgpt.com and sign in.
2Click your profile → Settings → Advanced and enable Developer Mode.
3Go to Settings → Connectors → Create, then paste the MCP URL above.
4Authorize Brandsearch when prompted.

Perplexity

Open Perplexity connectors

1Click the button below to open Perplexity's connector settings.
2Add a new MCP connector and paste the MCP server URL above.
3Authorize Brandsearch when prompted.

Cursor

1Open Cursor on your computer.
2Go to Settings → Tools & MCP (or Tools & Integrations).
3Click 'Add new MCP server' and paste the MCP URL above.
4Or edit ~/.cursor/mcp.json directly if you prefer.

Copilot

1Open VS Code with the GitHub Copilot extension installed.
2Press Ctrl+Shift+P (or Cmd+Shift+P on macOS) to open the Command Palette.
3Run 'MCP: Open User Configuration' and add Brandsearch using the MCP URL above.

Gemini

1Open ~/.gemini/settings.json in your editor.
2Add a new MCP server entry pointing to the MCP URL above.
3Restart the Gemini CLI for the change to take effect.

Windsurf

1Open Windsurf on your computer.
2Go to Settings → Cascade → MCP Servers.
3Click 'Manage Plugins' and add a new server with the MCP URL above.

Other MCP clients

Any client that supports remote MCP servers over HTTPS will work. Point it at the MCP URL above and follow the OAuth prompt.

For Claude Desktop, Cursor, Gemini, and any CLI that takes a JSON config, the entry looks like:

mcp config

{
  "mcpServers": {
    "brandsearch": {
      "url": "https://api.brandsearch.co/v1/mcp"
    }
  }
}

What you can ask

Once connected, just talk to your AI normally. The AI picks the right tool and arguments automatically.

Brands

“Find DTC beauty brands with 1M+ monthly visits and active Meta ads”

Meta ads

“Show me Nike's top 10 highest-spending Meta ads from 2025”

Brand profile

“Compare Allbirds and Nike — give me a summary of each brand”

TikTok posts

“Discover trending TikTok posts in the Pet Supplies niche”

Aggregates

“Get the dashboard for Glow Recipe's Meta ads — funnel, CTAs, top countries”

Products

“List the bestsellers for allbirds.com — title, price, and URL”

Account

“How many requests do I have left today?”

System

“Is today's Meta ads data ready yet?”

Media URLs are included in MCP responses. Image, video, thumbnail, and cover URLs are returned by every MCP tool the same way they are over HTTP — so your AI can show or link to ad creatives, product images, brand logos, and screenshots directly. Image URLs still expire 3 days after issue (see the ads and brands pages); re-run the tool to get fresh ones.

Available tools

The MCP server exposes 30+ tools. They map 1:1 to the API endpoints documented elsewhere in this site, with sensible defaults so the AI doesn't have to know every parameter.

Brands

Name
search_brands
Type
tool
Description
Filter brands by niche, monthly visits, active Meta ads, product count, apps, or technologies.
Name
discover_brands
Type
tool
Description
Random sample of trending brands. The right tool for "show me brands" / "give me examples".
Name
get_brand
Type
tool
Description
Full details for a brand by domain (e.g., nike.com).
Name
get_brand_by_url
Type
tool
Description
Full details for a brand by full website URL.
Name
get_brands_by_ids
Type
tool
Description
Batch fetch up to 100 brands by ID.
Name
lookup_brand
Type
tool
Description
Free-form search by brand name, domain, or social handle.
Name
lookup_brands_batch
Type
tool
Description
Batch look up to 50 identifiers in one call.
Name
query_brands
Type
tool
Description
Same as search_brands but takes a JSON filter body. Use when filters get long.
Name
get_brand_summary
Type
tool
Description
Profile + active Meta ads + products in a single call. Best for "tell me about X".
Name
get_brand_advertisers
Type
tool
Description
The Facebook/Meta pages running ads for a brand.

Ads & posts

Name
get_brand_ads
Type
tool
Description
Ads or posts for a specific brand on meta, tiktok, or instagram.
Name
search_meta_ads
Type
tool
Description
Search every Meta ad we track by copy, spend, reach, targeting, dates.
Name
search_tiktok_ads
Type
tool
Description
Search every TikTok post by description, engagement, duration.
Name
search_instagram_posts
Type
tool
Description
Search every Instagram post by caption, hashtags, engagement.
Name
query_meta_ads
Type
tool
Description
JSON-body counterpart of search_meta_ads.
Name
query_tiktok_ads
Type
tool
Description
JSON-body counterpart of search_tiktok_ads.
Name
query_instagram_posts
Type
tool
Description
JSON-body counterpart of search_instagram_posts.
Name
get_meta_ad
Type
tool
Description
Single Meta ad by ID.
Name
get_tiktok_ad
Type
tool
Description
Single TikTok post by ID.
Name
get_instagram_post
Type
tool
Description
Single Instagram post by ID.
Name
discover_meta_ads
Type
tool
Description
Random high-quality Meta ads.
Name
discover_tiktok_ads
Type
tool
Description
Random viral TikTok posts.
Name
discover_instagram_posts
Type
tool
Description
Random viral Instagram posts.

Aggregates (dashboards)

Name
get_brand_ads_aggregates
Type
tool
Description
Six-dimension Meta ads dashboard for a single brand (media mix, funnel, CTAs, demography, spend).
Name
get_brand_posts_aggregates
Type
tool
Description
TikTok or Instagram dashboard for a single brand.
Name
get_meta_ads_aggregates
Type
tool
Description
Meta ads dashboard across all brands matching a filter slice.
Name
get_tiktok_ads_aggregates
Type
tool
Description
TikTok dashboard across all brands.
Name
get_instagram_posts_aggregates
Type
tool
Description
Instagram dashboard across all brands.

Products & metadata

Name
get_products
Type
tool
Description
Products sold by a brand. product_type can be all, bestsellers, or latest.
Name
get_facet
Type
tool
Description
Canonical reference values (e.g., apps, technologies, niches, country-codes). Use when you need to know the valid values for a filter.
Name
get_usage
Type
tool
Description
Current API usage + remaining daily/monthly quota.
Name
get_system_freshness
Type
tool
Description
Latest timestamps for brands, Meta ads, TikTok posts, and Instagram posts. "Is today's data ready?"

Rate limits

The MCP endpoint has its own rate limits separate from the REST API. If your AI hits them, it will get a 429 too_many_requests response with a Retry-After header — the AI will typically retry after the suggested wait. Underlying tool calls still count against your normal API quota; check it any time with the get_usage tool.

Tips for getting good results

Use discover_brands for vague prompts. "Show me some cool DTC brands" should hit discover_brands, not search_brands with no filters.
Use get_brand_summary over chained calls. It bundles profile + ads + products in one round trip.
Use get_facet to validate values before filtering. If the AI is unsure whether Klaviyo is the canonical app name, it can call get_facet("apps") first.
Use query_* over search_* for long filter sets. JSON body avoids querystring length limits and keeps arrays clean.