Cursor peut appeler des outils externes via le Model Context Protocol, mais il est livré sans recherche en direct. L'ajout de Scavio MCP donne à Cursor un accès aux données de Google, YouTube, Amazon, Walmart, Reddit et TikTok dans n'importe quelle conversation, sans code API requis. Le point de terminaison MCP est https://mcp.scavio.dev/mcp et chaque appel d'outil coûte 0,005 $. Ce tutoriel le fait fonctionner en moins de 5 minutes.
Prérequis
- IDE Cursor installé
- Une clé API Scavio depuis scavio.dev
- Connaissance de base des paramètres de Cursor
Parcours
Étape 1: Obtenez votre clé API Scavio
Inscrivez-vous sur scavio.dev et copiez votre clé API depuis le tableau de bord.
Bash
# 1. Go to https://scavio.dev
# 2. Sign up (free tier: 250 credits/month, no card required)
# 3. Navigate to Dashboard > API Keys
# 4. Click "Create New Key"
# 5. Copy the key -- you'll need it in the next step
# Verify your key works:
curl -X POST https://api.scavio.dev/api/v1/search \
-H 'x-api-key: YOUR_KEY_HERE' \
-H 'Content-Type: application/json' \
-d '{"query": "test", "country_code": "us"}'
# You should see JSON with organic_resultsÉtape 2: Ajoutez Scavio MCP à la configuration de Cursor
Ouvrez les paramètres de Cursor et ajoutez la configuration du serveur MCP.
JSON
// File: ~/.cursor/mcp.json (create if it doesn't exist)
// On macOS: ~/Library/Application Support/Cursor/mcp.json
// On Windows: %APPDATA%\Cursor\mcp.json
{
"mcpServers": {
"scavio": {
"url": "https://mcp.scavio.dev/mcp",
"headers": {
"x-api-key": "YOUR_SCAVIO_API_KEY"
}
}
}
}Étape 3: Redémarrez Cursor et vérifiez la connexion
Rechargez Cursor pour qu'il prenne en compte le nouveau serveur MCP, puis testez-le.
Bash
# After saving mcp.json:
# 1. Restart Cursor (Cmd+Shift+P > "Reload Window" or quit and reopen)
# 2. Open a new Composer chat (Cmd+I)
# 3. Type: "Search Google for best python web framework 2026"
# 4. Cursor should call the Scavio search tool and return live results
# You can also verify in Cursor Settings > MCP
# The "scavio" server should show as "Connected"
# Available tools after connection:
# - search: Google, YouTube, Amazon, Walmart, Reddit
# - tiktok: profile, posts, video, comments, search, hashtags
# - extract: Pull structured data from any URLÉtape 4: Utilisez la recherche dans votre flux de travail de codage
Exemples d'invites qui exploitent les données de recherche en direct dans Cursor.
Bash
# Example prompts to try in Cursor Composer:
# Research-backed coding:
# "Search for the latest Next.js 15 API route syntax and update my route handler"
# Price comparison for docs:
# "Search Amazon for mechanical keyboards under $100 and create a comparison table component"
# Reddit-informed decisions:
# "Search Reddit for common complaints about Prisma ORM and suggest alternatives"
# TikTok data for marketing:
# "Get the TikTok profile for @username and summarize their engagement metrics"
# Multi-platform research:
# "Search Google for 'best auth library Node.js 2026' and also check Reddit for real user opinions"
# Each search tool call costs $0.005 (1 credit)
# Free tier: 250 credits/month = 250 searches
# Project plan: $30/month = 7,000 searchesExemple Python
Python
# MCP is a zero-code integration -- no Python needed.
# But if you want to test the same endpoint programmatically:
import os, requests
SH = {'x-api-key': os.environ['SCAVIO_API_KEY'], 'Content-Type': 'application/json'}
data = requests.post('https://api.scavio.dev/api/v1/search',
headers=SH, json={'query': 'cursor ide mcp setup', 'country_code': 'us'}).json()
for r in data.get('organic_results', [])[:3]:
print(f'{r["position"]}. {r["title"][:60]}')
print(f'Cost: $0.005')Exemple JavaScript
JavaScript
// MCP is a zero-code integration -- no JS needed.
// But if you want to test the same endpoint programmatically:
const SH = { 'x-api-key': process.env.SCAVIO_API_KEY, 'Content-Type': 'application/json' };
const data = await fetch('https://api.scavio.dev/api/v1/search', {
method: 'POST', headers: SH,
body: JSON.stringify({ query: 'cursor ide mcp setup', country_code: 'us' })
}).then(r => r.json());
(data.organic_results || []).slice(0, 3).forEach(r =>
console.log(`${r.position}. ${r.title.slice(0, 60)}`)
);
console.log('Cost: $0.005');Sortie attendue
JSON
# After adding MCP config and restarting Cursor:
Cursor Settings > MCP:
scavio: Connected
# In Composer chat:
User: Search Google for best python web framework 2026
Cursor: [Calling scavio.search with {query: 'best python web framework 2026'}]
Based on live search results:
1. FastAPI continues to lead for API development with async support
2. Django 5.1 added improved async views and type hints
3. Litestar emerged as a strong FastAPI alternative
...
Cost per search: $0.005
Free tier: 250 searches/month