Public ledger of every directory change — grade bumps, KYC re-tests, sponsorship activations, incident updates. We log it so visitors can verify our claim that the editorial firewall holds.
The list is the product; this is the receipt of how it changes.
xmr.club
{
"also_listed_at": {
"web3privacy": true
}
}{
"also_listed_at": {
"web3privacy": true
}
}{"provider":{"id":"llama-cpp","category":"ai","subcategory":"Local Runtime","name":"llama.cpp","tagline":"Reference C++ inference engine for open-source LLMs. Foundational layer beneath Ollama, LM Studio, Jan, KoboldCpp.","review":"**The reference C++ inference engine for open-source LLMs** — the project Ollama / LM Studio / Jan / KoboldCpp / llamafile all wrap, by the developer who invented the GGUF model format the open-weight ecosystem standardised on. Listed at **Grade A** because `llama.cpp` is the irreducible local-inference layer: no operator on the data path, MIT-licensed, single-binary install, zero account anywhere, native acceleration on CPU + NVIDIA CUDA + AMD ROCm + Apple Metal + Vulkan. The strongest privacy posture available in this directory, jointly with Ollama (which sits one wrapper layer above), and the right answer when you want **maximum control over quantisation, context window, batch size, and offload split** — the things Ollama abstracts away.\n\n**What it is.*{"provider":{"id":"llama-cpp","category":"ai","subcategory":"Local Runtime","name":"llama.cpp","tagline":"Reference C++ inference engine for open-source LLMs. Foundational layer beneath Ollama, LM Studio, Jan, KoboldCpp.","review":"**The reference C++ inference engine for open-source LLMs** — the project Ollama / LM Studio / Jan / KoboldCpp / llamafile all wrap, by the developer who invented the GGUF model format the open-weight ecosystem standardised on. Listed at **Grade A** because `llama.cpp` is the irreducible local-inference layer: no operator on the data path, MIT-licensed, single-binary install, zero account anywhere, native acceleration on CPU + NVIDIA CUDA + AMD ROCm + Apple Metal + Vulkan. The strongest privacy posture available in this directory, jointly with Ollama (which sits one wrapper layer above), and the right answer when you want **maximum control over quantisation, context window, batch size, and offload split** — the things Ollama abstracts away.\n\n**What it is.*{
"also_listed_at": {
"web3privacy": true
}
}{
"also_listed_at": {
"web3privacy": true
}
}{"provider":{"id":"llama-cpp","category":"ai","subcategory":"Local Runtime","name":"llama.cpp","tagline":"Reference C++ inference engine for open-source LLMs. Foundational layer beneath Ollama, LM Studio, Jan, KoboldCpp.","review":"**The reference C++ inference engine for open-source LLMs** — the project Ollama / LM Studio / Jan / KoboldCpp / llamafile all wrap, by the developer who invented the GGUF model format the open-weight ecosystem standardised on. Listed at **Grade A** because `llama.cpp` is the irreducible local-inference layer: no operator on the data path, MIT-licensed, single-binary install, zero account anywhere, native acceleration on CPU + NVIDIA CUDA + AMD ROCm + Apple Metal + Vulkan. The strongest privacy posture available in this directory, jointly with Ollama (which sits one wrapper layer above), and the right answer when you want **maximum control over quantisation, context window, batch size, and offload split** — the things Ollama abstracts away.\n\n**What it is.*