RepoRadar
Leave a Tip
Item detail

jundot/omlx

oMLX is an Apache-2.0 LLM inference server for Apple Silicon with a macOS menu-bar app, CLI, OpenAI-compatible API, continuous batching, model pinning, hot/cold KV cache tiers, SSD cache reuse, and optional MCP support. It targets local coding-agent use where prompt cache stability and easy model control matter.

Open source →Back to radar
Score7.9
Popularity67.0
Riskconditional
TierGold
Score breakdown
Usefulness8.0
Novelty7.0
Momentum7.0
Maturity7.6
Open-source/build8.4
Evidence7.2
Workflow potential9.0
Setup ease6.4

Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.

Why it matters

Useful for Mac-based local-AI users who want a more managed MLX server: start with a small model, keep it bound to localhost, and benchmark cache behavior against your current Ollama/MLX setup.

Who should use it

Apple Silicon userslocal LLM developerscoding-agent usershomelab inference testers

Who should skip it

Skip if the source link, docs, or setup requirements do not match your workflow.

Risk explanation

local inference endpoints should remain protected; macOS 15 and Apple Silicon requirements limit who can use it.

Evidence links

Closest alternatives / related signals

mlxapple-siliconlocal-aiopenai-compatiblemcp