Best Self-Hosted Speed Test Tools in 2026
Quick Picks
| Use Case | Best Choice | Why |
|---|---|---|
| On-demand speed tests | LibreSpeed | Run a test whenever you want, like Ookla but private |
| Automated monitoring | Speedtest Tracker | Scheduled tests with historical graphs and alerting |
| LAN performance testing | LibreSpeed | Tests internal network speed, not just WAN |
| ISP accountability | Speedtest Tracker | Logs historical data to prove service degradation |
Why Self-Host a Speed Test?
Third-party speed tests (Ookla, Fast.com) send your IP and test data to external servers. ISPs sometimes prioritize traffic to known speed test servers, giving artificially inflated results. A self-hosted speed test gives you:
- Accurate internal network measurements — test LAN speed between devices
- No data sent externally — your bandwidth data stays private
- ISP-honest results — test against your own server, not one ISPs optimize for
- Historical tracking — spot degradation over weeks or months
The Full Ranking
1. LibreSpeed — Best for On-Demand Testing
LibreSpeed is a self-hosted replacement for Ookla’s Speedtest. It runs entirely in your browser against your own server — no Flash, no Java, no external dependencies. Point your browser at it, click “Start,” and get download/upload speed, latency, and jitter measurements.
What makes it stand out: LibreSpeed tests the actual network path between your device and your server. This is invaluable for diagnosing LAN bottlenecks, testing Wi-Fi performance in different rooms, or verifying that your VPN isn’t throttling bandwidth.
Pros:
- Lightweight — runs on any hardware, including Raspberry Pi
- No external dependencies or accounts
- Tests internal and external network speed
- Multiple test modes (single stream, multi-stream)
- Clean, distraction-free UI
Cons:
- Manual testing only — no scheduling or historical data
- No alerting or notifications
- Basic results display (no graphs or trends)
Best for: Anyone who wants a private Ookla replacement. Especially useful for testing internal network performance.
Read our full guide: How to Self-Host LibreSpeed
2. Speedtest Tracker — Best for Automated Monitoring
Speedtest Tracker runs scheduled speed tests and stores the results in a database. You get historical graphs showing bandwidth over time, which is exactly what you need to prove to your ISP that their service is degrading at peak hours.
What makes it stand out: Automated scheduling with configurable intervals. Set it to test every 30 minutes and come back a week later to see trends. Threshold-based alerts notify you when speed drops below acceptable levels.
Pros:
- Automated scheduled tests (configurable intervals)
- Historical graphs and data retention
- Threshold-based alerting
- Dashboard with current and historical results
- Supports multiple speed test backends (Ookla CLI, LibreSpeed)
Cons:
- Heavier than LibreSpeed (requires database)
- Tests WAN speed by default (not internal LAN)
- Uses Ookla’s CLI under the hood (external servers)
Best for: Long-term ISP monitoring and proving service level violations. If you suspect your ISP throttles during certain hours, this provides the evidence.
Read our full guide: How to Self-Host Speedtest Tracker
Comparison Table
| Feature | LibreSpeed | Speedtest Tracker |
|---|---|---|
| Test Type | On-demand (manual) | Scheduled (automated) |
| Network Tested | Internal + External | External (Ookla servers) |
| Historical Data | No | Yes (graphs + database) |
| Alerting | No | Yes (threshold-based) |
| Database Required | No | Yes (SQLite or PostgreSQL) |
| RAM Usage | ~30 MB | ~200 MB |
| Docker Setup | One container | One container + DB |
| Multi-Server | Yes (test multiple endpoints) | No |
| API | Basic | REST API |
How We Evaluated
We prioritized accuracy, privacy, and ease of deployment. LibreSpeed wins for simplicity and internal network testing. Speedtest Tracker wins for automated monitoring with historical data. Most self-hosters benefit from running both — LibreSpeed for ad-hoc LAN testing and Speedtest Tracker for ISP monitoring.
Frequently Asked Questions
Is a self-hosted speed test more accurate than Ookla?
For internal network testing, yes. Ookla tests speed to external servers, which ISPs may optimize for. A self-hosted LibreSpeed instance tests the actual path between your devices and your server, revealing real LAN performance, Wi-Fi dead zones, and switch bottlenecks that Ookla never sees.
Can I test speed between two locations with LibreSpeed?
Yes. Deploy LibreSpeed on a remote VPS and run tests from your home network. This measures your real-world upload/download to that specific server — useful for testing VPN throughput, WireGuard performance, or Cloudflare Tunnel overhead.
How often should Speedtest Tracker run tests?
Every 30–60 minutes is a good default. More frequent testing can skew your bandwidth usage measurements and may trigger ISP throttling if they detect excessive speed test traffic. For ISP accountability, 30-minute intervals provide enough data points to prove time-of-day degradation patterns.
Do self-hosted speed tests use my bandwidth?
Yes. Each test transfers data between your device and the server. A typical LibreSpeed test uses 50–200 MB. Speedtest Tracker running every 30 minutes uses roughly 10–15 GB/month. Factor this into metered connections.
Can I run both LibreSpeed and Speedtest Tracker together?
Yes, and most self-hosters should. LibreSpeed handles ad-hoc LAN diagnostics (testing Wi-Fi in different rooms, comparing wired vs wireless). Speedtest Tracker handles automated ISP monitoring with historical graphs. Combined idle RAM is under 250 MB.
Will my ISP detect self-hosted speed tests?
They’ll see the traffic but can’t easily identify it as a speed test. Unlike Ookla (which connects to known test servers), LibreSpeed traffic looks like normal HTTP transfers to your own server. This means your ISP can’t selectively boost performance during tests.
Can LibreSpeed test 10 Gbps connections?
LibreSpeed can test high-speed connections but browser-based JavaScript has overhead. For 10G testing, the limiting factor is usually the browser’s WebSocket implementation and single-thread performance. LibreSpeed supports multi-stream testing which helps. For precise 10G benchmarking, dedicated tools like iperf3 are more reliable.
Related
Get self-hosting tips in your inbox
Get the Docker Compose configs, hardware picks, and setup shortcuts we don't put in articles. Weekly. No spam.
Comments