llama-swap
Proxy server that swaps local LLM models on demand for llama.cpp and similar backends.
Used alongside llama.cpp to swap local models on demand123, serving several models from a single endpoint4 and giving an Ollama-like experience.56 The most common caveat is that some former users switched to lemonade78, which one found easier to manage.9
Why people recommend it
llama.cpp companion · 3
Commenters pair it with llama.cpp for their local inference setups.
“llama.cpp + llama-swap”
“I use llama.cpp w/ llama-swap”
Show 1 more
“I just switched from Llama.cpp to Llama swap with the help of codex. It was great.”
Dynamic model swapping · 2
It swaps models dynamically and exposes multiple models on one endpoint.
“I use llama-swap to dynamically swap models.”
“llama-swap (running as a podman quadlet exposed via tailscale svc) allows running multiple models on a single endpoint.”
Ollama-like ease · 2
Commenters describe it as emulating the Ollama experience and as fairly easy to use.
“I use llama-swap to emulate the Ollama experience.”
“Use ollama or llama-swap which is fairly easy to use.”
Caveats and dissent
Lemonade alternative · 2
Some former users moved to lemonade, which one found easier to manage.
“I used llama swap for a while before leaning into lemonade.”
Manual model setup · 1
One user built custom scripts to create hardlinks so llama-swap could load GGUF files.
“populate a separate cache directory of hardlinks so llama-swap can load the gguf's into llama.cpp”
Often compared with
Who recommends it in their own words
Threads where it came up
- Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding? 3 mentions
- Kimi K2.7-Code: open-source coding model with better token efficiency 2 mentions
- Best LLM for every budget, updated daily 1 mention
- Meta Muse Glimmer – open weights 30B local coding model 1 mention
- Hands-On with the AMD Ryzen AI Halo 1 mention
- DGX Spark vs. Mac Studio and Halo 1 mention
Data: llama-swap.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 10
“llama.cpp + llama-swap”
“I use llama.cpp w/ llama-swap”
“I used llama swap for a while before leaning into lemonade.”
“previously I used llama-swap”
“I use llama-swap to dynamically swap models.”
“I use llama-swap to emulate the Ollama experience.”
“llama-swap (running as a podman quadlet exposed via tailscale svc) allows running multiple models on a single endpoint.”
“I just switched from Llama.cpp to Llama swap with the help of codex. It was great.”
“Use ollama or llama-swap which is fairly easy to use.”
“populate a separate cache directory of hardlinks so llama-swap can load the gguf's into llama.cpp”