How to Use a Proxy API

A proxy API lets an application request recently validated proxy endpoints instead of maintaining its own collection and health-checking pipeline.

1. Create an API key

Sign in to the IPPool console, activate a plan, and copy the generated API key. Treat it as a secret: keep it in an environment variable and never commit it to a public repository.

export IPPOOL_API_KEY="your-api-key"

2. Request proxies

Call the extraction endpoint over HTTPS. The exact filters available depend on the active plan. Start with a small batch and inspect the response before increasing volume.

curl --fail --silent \
  "https://ippool.cc/api/proxy?key=${IPPOOL_API_KEY}&protocol=http&num=10"

See the complete API documentation for current parameter names, response formats, country filtering, and AWS filtering.

The protocol parameter selects one of three independently validated modes — http for plain forwarding, https for CONNECT tunnelling, and tls for an encrypted connection to the proxy itself. Omit it to receive every enabled mode; see proxy modes explained for the details.

3. Validate and rotate responsibly

Even a recently checked proxy can become unavailable between validation and use. Your client should set connection and request timeouts, retry only idempotent operations, and rotate to another endpoint after a network failure or a clear proxy error.

• Use bounded retries with exponential backoff.
• Do not retry authentication or application errors blindly.
• Respect target-site terms, robots rules, and rate limits.
• Never log complete API keys.

4. Choose rotation or sticky sessions

Rotation is useful for independent requests. A sticky proxy is better when a website session, cookie jar, or multi-step workflow expects the same source IP. Keep proxy assignment separate from business logic so it can be replaced safely after failure.

5. Monitor outcomes

Track connection failures, TLS failures, response status, latency, and successful business outcomes separately. A connection can succeed while the target response is still unusable. These measurements help select sensible timeouts and retry thresholds.