How to Choose Clash Proxy Nodes:
A Complete Selection & Testing Guide

Three Key Metrics

Not all proxy nodes perform the same. Even different nodes from the same provider can vary significantly in speed, stability, and suitability for different tasks. When evaluating nodes, focus on these three dimensions:

• Latency: Round-trip time for data. Lower is better — under 100ms is good; under 50ms is excellent. Latency directly affects how snappy web pages and interactive apps feel. High latency causes noticeable sluggishness even if bandwidth is high.
• Packet Loss: Percentage of dropped packets. Directly affects connection stability — especially important for gaming, video calls, and real-time applications. Any packet loss above 1% is noticeable; above 5% causes significant degradation.
• Bandwidth: Maximum throughput of the node — determines the ceiling for video streaming and file download speeds. A high-latency node with great bandwidth is better for downloads; a low-latency node with moderate bandwidth works best for interactive use.

Of the three metrics, latency determines how snappy web pages and interactive apps feel; packet loss determines connection stability; bandwidth determines how well large downloads and high-bitrate video streams perform. The right balance depends on your use case — and often the best approach is to test before committing to a node for critical tasks.

Testing Nodes with Clash's Built-In Tool

Clash Verge Rev includes a built-in latency test — no third-party tools needed:

1. Click the "Proxies" icon in the left sidebar to open the node management page.
2. Find the speed test icon (typically a lightning bolt or refresh icon) in the top-right corner of the node list, and click it.
3. The client sends concurrent test requests to all nodes. This usually completes in 10–30 seconds depending on how many nodes are in your subscription.
4. After the test, each node displays its latency: green = low (good), yellow = medium, red = high or timed out. Timed-out nodes are unreachable and should be avoided.
5. Select a green node with the lowest latency and click to switch. The change takes effect immediately without disconnecting your current sessions.

Geography and Latency

Physical distance is one of the biggest factors affecting latency. Light traveling through fiber takes time — and with routing, processing, and hops added on top, the actual latency is always higher than the theoretical minimum. As a rule of thumb, nodes that are geographically closer to you will typically offer lower latency.

As a general guideline, choosing nodes geographically closer to you typically gives lower latency. For example:

• Users in North America: US West Coast or East Coast nodes offer the lowest latency for most use cases. US nodes are also ideal for accessing US-region streaming libraries (Netflix US, Disney+, HBO Max, etc.).
• Users in Europe: European nodes (UK, Germany, Netherlands, France) typically provide the lowest latency. UK and US nodes are both good for English-language content.
• Users in East/Southeast Asia: Nodes in Singapore, Japan, Hong Kong, or Taiwan often offer the best balance of latency and bandwidth for this region. Singapore nodes are especially popular due to excellent international connectivity.
• All users: For specific streaming platforms or services, always pick a node in the region where that service has the broadest content library — geographic location of the IP matters more than latency for streaming access.

Choosing Nodes by Use Case

Everyday Browsing & Social Apps

Prioritize the lowest-latency node available. For these use cases, stability matters more than raw latency numbers — a 40ms node that never drops is far better than a 20ms node that disconnects twice a day. Test a node over time rather than just relying on a single speed test snapshot. If your provider has a "Premium" or "Stable" tier, these nodes usually have better uptime guarantees.

Streaming Video (Netflix, Disney+, YouTube, Hulu, etc.)

Streaming platforms check the geographic location of the connecting IP and restrict access based on licensing agreements. When choosing a streaming node, first verify it can unlock the target platform, then consider latency and bandwidth. Providers usually label nodes with "Netflix", "Streaming", "Media", or similar tags in their node names or documentation. Bandwidth is critical for streaming: 4K video requires at least 25 Mbps of stable throughput, and 1080p requires around 5–8 Mbps. If you experience buffering, check your actual download speed on the node before blaming the proxy.

Online Gaming

Gaming is the most sensitive to both latency and packet loss. High latency causes input lag and stuttering; any packet loss can cause "teleporting," rubberbanding, or disconnections. When choosing a gaming node:

• Prioritize the lowest latency (ideally under 80ms to the game server, not just to the proxy).
• Verify packet loss is 0% using the built-in test or an in-game network diagnostic tool.
• Enable TUN mode so UDP game traffic is correctly routed through the proxy — System Proxy alone won't cover most game connections.
• Choose a node geographically close to the game server, not just close to you. For example, a JP node may give better performance in Japanese game servers even if you're in Southeast Asia.

Large File Downloads

Downloads are mainly about bandwidth. Look for nodes marked "high-speed", "unlimited", or "10Gbps" by your provider. Monitor your actual download speed on the node — if it's significantly lower than expected, try a different node in the same region. Some providers throttle speeds on shared nodes during peak hours; premium or dedicated nodes usually offer more consistent throughput.

Using the Auto-Select Policy Group

Clash config files support a url-test policy group that automatically tests all included nodes and switches to the lowest-latency one at a set interval. This is ideal for users who don't want to manually switch nodes:

proxy-groups:
  - name: Auto
    type: url-test
    proxies:
      - US01
      - JP01
      - SG01
    url: "http://www.gstatic.com/generate_204"
    interval: 300

interval: 300 means Clash re-tests every 5 minutes and switches if a faster node is found. Most provider subscriptions already include an "Auto" group configured this way — you can simply select it in the Proxies tab without any manual configuration.

When to Override Auto-Select

Auto-select is convenient, but there are cases where manual node selection is better:

• Streaming: Auto-select picks the fastest node, but the fastest node may not unlock the streaming platform you need. Manually pick a node labeled for your target service.
• Stability: If you're in a long video call or multiplayer game session, staying on a slightly slower but highly stable node is better than having Clash switch mid-session.
• Testing: When diagnosing issues, manually pin to a specific node to isolate whether the problem is the node or your rules.

For more details on proxy setup, visit our Help Center or check the Rule-Based Routing guide to optimize which traffic goes through your chosen node.