LocalAI · Issues· 181 open

GPU RAM detection with incorrect VRAM value

localai cannot run the recommended model

Docs: gfx1151 env vars are documented as "set automatically in the ROCm/hipblas image" but the image sets none of them

User-controlled Multi-Agent Conversation (Human-in-the-loop)

reasoning extraction only happens at end of stream

/api/ps and /api/tags report size and size_vram as a hardcoded 0

docker-compose.yaml default model phi-2 no longer exists in the gallery

Socket activation not working

Expose per-model memory usage (the per-process VRAM walk already computes it, then discards the split)

vllm-omni backend (cuda13) fails to load: editable install points at build path, and unpinned vllm-omni main no longer matches backend.py

Tool call arguments are empty on /responses but correct on /chat/completions (same model, same config)

Instella-MoE models in the LocalAI gallery is not supported by any backends

audio-cpp using a fish-audio doesn't work with referance audio/text

Gallery model audio-cpp-sortformer-diarization fails to install: SHA mismatch, reproducible across retries

Agent pool Postgres DSN is logged in cleartext, including the password, while the auth DSN is redacted