Lemonade
AMD-backed open-source local server for running LLMs, image, and speech models on PCs.
HN commenters recommend Lemonade for running local models on AMD hardware, especially Strix Halo desktops,1234 with ROCm, Vulkan and NPU backends plus text, image and speech generation in one easy install.567 Caveats include models defaulting to small 4K context windows and lagging behind upstream engine releases,8 and one user prefers its standalone llama.cpp ROCm build to the combined server.9
Why people recommend it
AMD/Strix Halo support · 3
Commenters recommend it as the go-to local AI stack for AMD hardware, especially Strix Halo.
“Highly recommend lemonade server if you have a strix halo desktop.”
“Really recommend this for AMD hardware!”
Show 1 more
“Check out the officially supported project Lemonade[0] by AMD.”
Many backends · 3
It bundles llama.cpp, vLLM, ComfyUI and other backends covering text, image, TTS and STT across ROCm, Vulkan, CPU, GPU and NPU.
“Really recommend this for AMD hardware!”
“Check out the officially supported project Lemonade[0] by AMD.”
Show 1 more
“Been running lemonade for some time on my Strix Halo box.”
Easy install · 3
Users note quick installs, ROCm handled through the app, and rpm, deb and AppImage packages.
“Check out the officially supported project Lemonade[0] by AMD.”
“I use Lemonade and have only the ROCm installed through it”
Show 1 more
“Neat, they have rpm, deb, and a companion AppImage desktop app[1]!”
Generation speed · 2
Strix Halo users report around 50 to 60 tokens per second on large models.
“Highly recommend lemonade server if you have a strix halo desktop.”
“Been running lemonade for some time on my Strix Halo box.”
Active development · 2
Commenters praise a pragmatic development pace and a much-improved UI.
“Really recommend this for AMD hardware!”
“I used llama swap for a while before leaning into lemonade.”
Compatible APIs · 1
OpenAI- and Ollama-compatible endpoints let it work with VSCode Copilot and Open WebUI.
“Really recommend this for AMD hardware!”
Caveats and dissent
Small default context · 1
Most models default to very small 4K context windows.
“I used llama swap for a while before leaning into lemonade.”
Upstream lag · 2
Bundled engines lag upstream releases, its llama.cpp is a fork, and vLLM support is experimental.
“I used llama swap for a while before leaning into lemonade.”
“I use Lemonade and have only the ROCm installed through it”
Combined server · 1
One user dislikes the combined server and uses only its llama.cpp ROCm build instead.
“Been running lemonade for some time on my Strix Halo box.”
Limited NPU benefit · 1
The NPU targets small always-on models, so it is not a big win for standard chatbot use.
“Been running lemonade for some time on my Strix Halo box.”
Unclear positioning · 1
One commenter could not tell what it is from the website or whether it supports Macs.
“An alternative to LM Studio? Does it support MLX or metal on Macs?”
Often compared with
Who recommends it in their own words
Threads where it came up
- Lemonade by AMD: a fast and open source local LLM server using GPU and NPU 4 mentions
- Hands-On with the AMD Ryzen AI Halo 2 mentions
- Show HN: Share your AI Setup, Learn from others 1 mention
- New Framework Desktop Option with AMD Ryzen AI Max+ Pro 495 and 192GB Memory 1 mention
- Qwen 3.8 Max Preview 1 mention
- DGX Spark vs. Mac Studio and Halo 1 mention
Also written as lemonade-server, Lemonade Server. Data: lemonade.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 13
“Using: LM Studio (primary), Lemonade”
“I use Lemonade and have only the ROCm installed through it”
“I used lemonade with Vulkan and the UD-Q8_x_x model from HF.”
“I used llama swap for a while before leaning into lemonade.”
“Highly recommend lemonade server if you have a strix halo desktop.”
“lemonade is just an easier to manage system”
“I haven't tried it with https://lemonade-server.ai/ yet but I just might give it a shot.”
“Check out the officially supported project Lemonade[0] by AMD.”
“Been running lemonade for some time on my Strix Halo box.”
“Really recommend this for AMD hardware!”
“An alternative to LM Studio? Does it support MLX or metal on Macs?”
“Neat, they have rpm, deb, and a companion AppImage desktop app[1]!”
Does Qwen3 Coder do a good job invoking its tools as appropriate for you? Under continue.dev at least, I've found I need to remind it constantly.