HProxy free-proxy-list: a continuously verified proxy feed with real metadata

What this repository actually gives you

The HProxy free-proxy-list solves a mundane but real problem: collecting free public proxies from many scattered sources and normalizing them into one place. The README claims over 671,906 proxies tracked all-time, with 42,168 live at the time of writing. That number is a snapshot, not a promise. The value is not the raw count; it is the attached metadata. Each proxy in all.json and all.csv carries protocol, anonymity grade, country, city, latency in milliseconds, and uptime over 24 hours and 7 days. Most free proxy lists hand you a bare ip:port and nothing else. This one gives you enough to filter before you even make a request. The target audience is developers who need a quick pool of candidates for scraping, testing, or geolocation checks, and who do not want to assemble their own aggregator.

How verification works, as far as the documentation shows

The README describes a pipeline: candidates are pulled from over 100 public sources every few minutes, then tested through HTTP, HTTPS, SOCKS4, and SOCKS5 around the clock. Each proxy is labeled with the protocols it actually answered on, not just the ones it was listed under. That distinction matters because a proxy advertised as SOCKS5 might only respond on HTTP. The repository also grades anonymity as transparent, anonymous, or elite, based on whether the proxy leaks your real IP. Latency and uptime are measured, not guessed. The published files hold proxies confirmed working within the last 48 hours. That freshness window is short, which is good for accuracy but means the list changes constantly. The README does not explain how often the files are regenerated beyond saying the list is refreshed every few minutes. The exact testing methodology, such as what request is sent or how many retries, is not documented in the material.

Getting the data: raw files, CDN, and code snippets

You do not need to clone the repository to use it. The README points to raw GitHub URLs for every file. For example, curl https://raw.githubusercontent.com/hproxy-com/free-proxy-list/main/all.txt returns every proxy as ip:port lines. The by-country folder has per-country files like US.txt. For richer data, all.json and all.csv include protocol, anonymity, country, city, latency, and uptime. The README also mentions a jsDelivr CDN mirror at https://cdn.jsdelivr.net/gh/hproxy-com/free-proxy-list@main/all.json to avoid rate limits. There is a Python example using requests: fetch all.txt, split by whitespace, pick a random proxy, and pass it as an http and https proxy to a request. A curl one-liner picks the first line of http.txt and uses it with -x. A Node.js snippet shows fetching the same txt file. The files are raw endpoints, so you can treat them as a simple API without authentication.

Filtering options and what they mean in practice

The repository ships pre-filtered files: http.txt, https.txt, socks4.txt, socks5.txt, elite.txt, anonymous.txt, fast.txt, and by-country files for 133 countries. The elite file contains high-anonymity proxies, the anonymous file contains the anonymous tier, and fast.txt lists proxies with latency under 2 seconds. The counts in the README give a sense of scale: 2,868 HTTP, 2,467 HTTPS, 2,683 SOCKS4, 3,352 SOCKS5, 4,649 elite, and 4,012 fast. These numbers are snapshots from the README, not current values. The filtering is useful, but you should not assume that elite means safe. The anonymity grade only indicates whether the proxy leaks your IP in the test the project runs. It says nothing about whether the proxy operator is logging your traffic. For sensitive requests, a free proxy from any list is a risk.

The obvious limitation: free proxies are ephemeral and untrusted

The biggest limitation is inherent to free proxies, not this repository. The README itself warns to treat them as public and likely unsafe, though the text is truncated before the full warning. The 48-hour freshness window means a proxy that works today may be dead tomorrow. The pool of 42,168 live proxies is large, but many will be slow or unreliable for real workloads. The uptime fields help, but they only reflect the project's own tests, not your specific target site. Another limitation is that the raw txt files give you no metadata, so if you want to filter by country or anonymity, you must parse all.json or all.csv. The README does not document the exact schema of those files beyond listing field names. If you need a stable, guaranteed proxy for a production system, this list is the wrong tool. You would be better off with a paid residential proxy service, which the README mentions as an option on hproxy.com.

A real alternative: running your own proxy checker

Instead of relying on this aggregated list, you could build your own proxy checker. The approach differs fundamentally: this repository gives you pre-verified candidates, but you have no control over the verification frequency or methodology. A self-hosted checker would pull from the same public sources, test proxies against your own target endpoints, and maintain your own live pool. Tools like Scrapy with its built-in proxy middleware, or a simple Python script using aiohttp to test many proxies concurrently, would give you full control. The trade-off is maintenance: you must handle source scraping, deduplication, protocol detection, and failure handling yourself. This repository saves you that work, but you inherit its verification schedule and its blind spots. If your use case requires testing against a specific site that blocks certain proxy types, running your own checker is the only way to know for sure.

Maintenance and licence considerations

The repository is under the MIT license, which means you can use the code and data freely with attribution. There are no recent releases listed, and the last push date is unknown, but the README shows an update date of 29.08.2026, indicating active maintenance at the time of writing. The project is not archived, so it is likely still being updated. The maintenance cost for you as a user is low: you do not need to install anything, just fetch files. However, you must handle the constant churn. If you cache the list, you need a refresh strategy. The README suggests using the CDN mirror to avoid rate limits, which implies that raw GitHub URLs could be rate-limited if you poll too frequently. The Apify actor is offered as an alternative with no signup, but it is a separate service. The licence does not cover the proxies themselves; they are third-party resources with their own unknown terms.

Editorial conclusion

Adopt this repository if you need a quick, no-signup source of many proxy candidates for low-stakes tasks like scraping public pages or testing proxy logic. Skip it if you need reliable proxies for production traffic, because free proxies are ephemeral and untrusted by nature. Before using any entry, verify the proxy yourself against your target, check the anonymity field in all.json, and never send credentials through a proxy from this list. The repository is a convenient aggregator, not a guarantee of safety or uptime.

Frequently asked questions

Official sources

1. Official documentation
2. Official README
3. Project repository