Claude MCP setup
Search Facebook ads directly from Claude
Give Claude access to Spresh ad intelligence so it can research competitors, compare hooks, summarize creative strategy, and build briefs from real Meta ads.
Example agent request
MCP connectedFind active protein bar ads running in the US. Group them by hook, audience, offer, and creative format. Show which angles appear most often.
Get started in seconds
Copy one prompt. Let your agent install the MCP.
Pick your AI agent, copy the setup prompt, paste it into the agent, and add your Spresh API key when asked. The flow is built for speed.
Claude Prompt based installerPaste this into Claude
Install the Spresh Facebook Ad Library MCP in Claude. Use npx -y spreshapp-mcp, ask me for my Spresh API key, add it as SPRESHAPP_API_KEY, then verify the server is connected and show me three example prompts for competitor ad research.
Manual config fallback
{
"mcpServers": {
"spreshapp-ad-library": {
"command": "npx",
"args": ["-y", "spreshapp-mcp"],
"env": {
"SPRESHAPP_API_KEY": "your_api_key_here"
}
}
}
}Setup checklist
- 1 Copy the Claude install prompt.
- 2 Paste it into Claude Desktop or Claude Code.
- 3 Add your Spresh API key when Claude asks, then restart the MCP server.
Try after setup
“Find active ads from Athletic Greens and group them by hook.”
“Show skincare ads running in the US for more than 60 days.”
“Analyze the offer structure across these competitor ads.”
MCP tools
What your agent can call
search_ads
Search active and inactive Facebook ads by keyword, country, format, page, and niche.
search_brands
Find Meta pages for brands and competitors before pulling their active ad library.
get_brand_ads
List active ads for a brand with creative URLs, copy, landing pages, and metadata.
get_ad_details
Fetch ad transparency data, creative fields, audience regions, and ad level details.
analyze_creatives
Turn ad examples into hooks, angles, personas, offers, objections, and creative briefs.
Workflows
Built for ad intelligence work, not toy demos
Competitor ad research
Ask your agent to find active ads from a brand, group them by hook, and surface the creative patterns worth testing.
Creative strategy briefs
Turn ad examples into personas, offers, objections, landing page notes, and new creative angles for your next sprint.
Niche discovery
Search by market, country, format, and topic to understand what advertisers are currently running in a category.
Agent pages
Install guides for every agent
Each guide has agent-specific setup prompts, examples, and SEO focused installation details.
FAQ
Questions before you connect
Does the Claude Ad Library MCP work with Claude Desktop?
Yes. You can use the JSON configuration with Claude Desktop or use the Claude Code command flow if you prefer a terminal based setup.
Can Claude analyze ad creative after search?
Yes. Claude can use the ad data returned by Spresh to summarize hooks, personas, formats, offers, claims, and creative angles.
What is an Ad Library MCP?
An Ad Library MCP connects an AI agent to Facebook and Meta ad research tools through the Model Context Protocol. Instead of switching tabs, the agent can search ads, inspect competitor creatives, and summarize patterns inside the chat.
Do I need to code to use the Spresh Ad Library MCP?
No. Most users can copy the install prompt for their agent, paste it into the assistant, add their API key, and start asking ad research questions.
What can I ask after installing it?
You can ask for competitor ads, long running creatives, hook analysis, country specific ad examples, landing page patterns, and creative brief ideas based on real Facebook ad data.
Ready
Connect your agent to live ad intelligence
Start with free credits, install the MCP, and ask your agent to research real ads instead of guessing from stale swipe files.
Get started free