Search results

bug: qmd query fails with CUDA error ggml-cuda.cu:98 despite GPU being available

bugSomething isn't workingSomething isn't working

requires triageRequires triagingRequires triaging

Status: Open.

Bug

#598 In withcatai/node-llama-cpp;

· wzgrx opened on Apr 20, 2026

spawnCommand pipes stdin/stdout when used inside MCP server, breaking JSON-RPC transport

Status: Open.

#596 In withcatai/node-llama-cpp;

· tosoyn opened on Apr 16, 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

feat: gemma4 model architecture

new featureNew feature or requestNew feature or request

Status: Open.

Feature

#594 In withcatai/node-llama-cpp;

· ittaboba opened on Apr 10, 2026

bug: Compilation fails on 32-bit platforms: error: use of overloaded operator '[]' is ambiguous (with operand types 'Napi::Uint32Array' (aka 'TypedArrayOf<unsigned int>') and 'int')

bugSomething isn't workingSomething isn't working

requires triageRequires triagingRequires triaging

Status: Open.

Bug

#593 In withcatai/node-llama-cpp;

· yurivict opened on Apr 8, 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

bug: commonJS headerReadResultPromisable instanceof Promise is false

bugSomething isn't workingSomething isn't working

requires triageRequires triagingRequires triaging

Status: Open.

Bug

#563 In withcatai/node-llama-cpp;

· jrk94 opened on Feb 25, 2026