minimlx — local LLMs on Apple Silicon

raz@spacefrigate ~/projects/minimlx % minimlx chat loading qwen38-27b-mtplx…

minimlx chat · qwen38-27b-mtplx · tools: 7 · bash: on Enter submits · Alt+Enter newline · /help · Ctrl-C stops gen · Ctrl-D exits

    › what is minimlx?
    thinking…
    Short answer wanted. Say what it is, where it runs, and that it stays on the machine.
    I run Gemma 4, Qwen 3.6 and Qwen 3.8 entirely on your own Mac, through Apple's MLX — chat, agentic coding, voice, and an Anthropic-compatible server for Claude Code. Nothing leaves the device.
      68 tok ·  57.0 tok/s · thought 84 chars
  

[ view on github ] [ see the numbers ]

› minimlx --help

chat: Interactive REPL — streaming output, reasoning, and a local tool loop: read, write, edit, grep, glob, ls, bash.
ask: One-shot prompt. Bare text works too, so minimlx "explain crc32" is enough.
code: One-shot programming task against your working tree, with the same tools.
conversation: Voice: speak → transcribe → generate → speak back, all on-device.
serve: Anthropic-compatible HTTP server, so Claude Code can run against a local model.
pull · ls · rm: Model and cache management, with aliases for everything in the library.

› minimlx models # measured: M5 Max 128 GB, cold, 300 tok @ temp 0.7

alias	build	decode	prefill	peak
qwen38-27b-mtplx	27B · 4/8-bit · MTP head	57 tok/s	352 tok/s	21.0 GB
qwen38-9b-4bit	9B · 4-bit	98 tok/s	358 tok/s	5.3 GB
qwen38-9b	9B · 8-bit	58 tok/s	276 tok/s	9.7 GB
qwen38-27b-4bit	27B · 4-bit	32 tok/s	234 tok/s	15.5 GB
qwen38-27b	27B · 8-bit	16 tok/s	159 tok/s	28.9 GB

Sustained load roughly halves all of it — an M-series laptop throttles well before the model does. The answer above streamed at the top row's rate.

› Read calc.py and say what is wrong with the add function. Do not edit anything.

thinking… The user is asking me to read calc.py and tell them what's wrong with the add function. No edits needed. [called tool] read_file({"path": "calc.py"}) def add(a, b): return a - b The add function is wrong because it subtracts instead of adding: def add(a, b): return a - b # should be: return a + b So add(2, 3) returns -1 instead of 5. The fix would be changing - to +, but as requested I haven't edited anything. 94 tok · 25.6 tok/s · thought 51 chars

› install

$ pip install "minimlx[mtplx] @ git+https://github.com/bravenewxyz/minimlx"

# then
$ minimlx chat
$ minimlx code "add a --json flag"

# or point Claude Code at your own hardware
$ minimlx serve
$ export ANTHROPIC_BASE_URL=http://127.0.0.1:1234

macOS on Apple Silicon · Python 3.11+ · MIT · github.com/bravenewxyz/minimlx