All issues

Search results

feat: MTP support

new featureNew feature or requestNew feature or request

Status: Open.

Feature

#642 In withcatai/node-llama-cpp;

· JoshuaWise opened on Aug 18, 2026

·  

v4.0.0

testBindingBinary fails on Windows: forked child doesn't send 'ready' with ignore stdio (Vulkan prebuilt falls back to source build)

Status: Open.

#638 In withcatai/node-llama-cpp;

· luluxiang06 opened on Aug 7, 2026

bug: gemma4 infinitely repeating a token randomly...

bugSomething isn't workingSomething isn't working

requires triageRequires triagingRequires triaging

Status: Open.

Bug

#627 In withcatai/node-llama-cpp;

· onzag opened on Jul 7, 2026

Gemma wrapper silently drops systemPrompt — system instructions never reach the model

Status: Open.

#624 In withcatai/node-llama-cpp;

· aschulman42-cell opened on Jul 2, 2026

bug: loadBackends(backendsPath) skipped when buildGpu === false, silently drops custom GPU backends compiled via NODE_LLAMA_CPP_CMAKE_OPTION_*

bugSomething isn't workingSomething isn't working

requires triageRequires triagingRequires triaging

Status: Open.

Bug

#599 In withcatai/node-llama-cpp;

· Zighy opened on Apr 25, 2026

Metal: embed() hangs sporadically on Apple M5 Max (~10-30% of calls)

Status: Open.

#595 In withcatai/node-llama-cpp;

· jowitte opened on Apr 14, 2026

Segmentation fault during embedBatch on ARM64 Linux (aarch64)

Status: Open.

#590 In withcatai/node-llama-cpp;

· yeHHH1g opened on Apr 3, 2026

feat: function calling across models

new featureNew feature or requestNew feature or request

requires triageRequires triagingRequires triaging

Status: Open.

Feature

#589 In withcatai/node-llama-cpp;

· sroussey opened on Apr 2, 2026

feat: git cloning should verify the cryptograhic hash

new featureNew feature or requestNew feature or request

requires triageRequires triagingRequires triaging

Status: Open.

Feature

#587 In withcatai/node-llama-cpp;

· yurivict opened on Mar 30, 2026

feat: Support native logprobs in LlamaChatSession.prompt() to avoid double inference

new featureNew feature or requestNew feature or request

Status: Open.

Feature

#584 In withcatai/node-llama-cpp;

· steve02081504 opened on Mar 22, 2026

·  

v4.0.0

Windows CUDA: cuMemAddressReserve failure in VMM pool causes hard abort (GGML_CUDA_NO_VMM workaround)

Status: Open.

#580 In withcatai/node-llama-cpp;

· enceos opened on Mar 15, 2026

pnpm --frozen-lockfile skips @node-llama-cpp/linux-x64 when lockfile generated on macOS

Status: Open.

#567 In withcatai/node-llama-cpp;

· IRIS-KELLER opened on Feb 28, 2026