Before
Unshaped circuit
- Idle latency
- 18 ms
- Loaded latency, download
- +91 ms
- Loaded latency, upload
- +142 ms
Poor
Keep subscriber connections responsive under load with intelligent, topology-aware traffic shaping.
Deploy LibreQoSSee how it works


LibreQoS models the network you actually built. Subscribers get per-flow CAKE fairness while shared capacity at APs, PONs, towers, sites, and backhaul stays under explicit hierarchical control. When one branch is constrained, congestion is contained there instead of degrading everyone upstream.
LibreQoS uses explicit queue control and CAKE-backed fairness instead of opaque congestion logic whose behavior is hard to predict under load.
Deterministic queue control02Topology-aware enforcement keeps shaping aligned to the AP, site, OLT, backhaul, or shared segment that is congested.
Topology-aware control03Operators can go from a network-wide issue into site, circuit, and subscriber-level evidence with RTT, retransmits, QoO, and Insight drill-down.
Subscriber-level proof fasterLibreQoS is purpose-built for wireless and fiber operators — shaping at real AP, site, OLT, and backhaul choke points with subscriber-level evidence. The same hierarchy runs dense shared-capacity environments, from event venues to vessels.
Reduce peak-hour complaints and keep shaping aligned to real AP, site, and backhaul bottlenecks.
Backhaul-aware shapingGive support teams stronger evidence about congestion, upstream trouble, and subscriber-side issues.
Subscriber evidence fasterGive managed-service teams a repeatable way to allocate shared capacity, compare locations, and investigate guest complaints across events, venues, hotels, vessels, and aircraft.
Managed WiFi Operators“We can afford to deploy a box at every step of the network to manage congestion at very little cost.”
Louis ElliottCEO, Streamline Internet
“LibreQoS let subscribers share higher plan speeds fairly and made the service feel much faster during prime time.”
Daniel DensonDenson Technologies
The public Internet Quality Test measures what subscribers feel on a loaded circuit — idle and loaded latency, jitter, packet loss, throughput, and per-application QoO — and grades the connection from A+ to F. Run it on a busy-hour circuit, deploy LibreQoS, then run it again.
Before
Poor
After
Excellent
Illustrative before/after readings for a bufferbloat-limited circuit. Run the test on your own network for the numbers that matter.
LibreQoS uses CAKE to keep busy networks responsive. Instead of letting downloads and streaming bursts crowd out everything else, it helps gaming, voice, and other interactive traffic stay usable under load.
Basic queueing
A basic queue sends packets in the order they arrive. When download bursts fill the line, interactive traffic has to wait, and the network starts to feel sluggish.
Queue dynamics
What subscribers feel
LibreQoS with CAKE
LibreQoS uses CAKE to keep interactive traffic moving even when the network is busy, so calls, gaming, and web use stay more responsive.
Queue dynamics
What subscribers feel
LibreQoS lets operators watch throughput, queue build-up, and RTT change live on an individual circuit, not just in after-the-fact charts.
See how LibreQoS fits into your stack and how to roll it out cleanly.
Review hardware fit, install flow, validation, and rollout guidance in one place.
Meet the team, see the open-source foundation, and understand the company behind LibreQoS.
Open the docs, support channels, GitHub project, and operator resources that help teams validate rollout cleanly.
The shaping core is open source and self-serve. Paid plans add support, API, and Insight — with volume brackets that drop the rate as you grow.
Shaping Core
The packaged build includes topology-aware shaping and the local console for up to 1,000 valid mapped circuits. Install on your own hardware.
LibreQoS Local
A production license for larger mapped-circuit deployments, priced per subscriber with regional rates available.
Insight
History from 15 minutes to one year, ASN analysis, and centralized views across one or many shapers. Volume brackets reduce the rate as you grow.
US list rates shown. Volume brackets reduce the per-subscriber rate as you grow; regional pricing available.
Run 1 CLI line, open the WebUI, complete setup, and validate against a known-good pilot before production cutover.