← Back to Blog
How-to

Run Claude Code with LM Studio: a local LLM on Apple Silicon

Connect Claude Code to LM Studio's Anthropic-compatible endpoint and run a local model on Apple Silicon: setup, model choice, context length, and the errors you will actually meet.

Claude Code can talk to any server that speaks Anthropic's Messages API, and since version 0.4.1 (January 2026) LM Studio serves exactly that on localhost:1234. Two environment variables point Claude Code at it; no proxy, no provider file, no translation layer. This guide covers the setup on a Mac with Apple Silicon, how to choose a model and context length that actually work with an agent, and the failure modes I ran into. It was first written in April 2026 and revised in September 2026 for the current LM Studio and Claude Code behaviour.

I use this for the same reason many scientists will: unpublished sequence data and manuscripts should not leave the machine, and a local model costs nothing per token. The trade-off is capability. A 20–30B local model is a competent assistant for scripts, refactoring, and documentation; it is not a substitute for the frontier models on large multi-file tasks.

Prerequisites

  • A Mac with Apple Silicon and enough unified memory for the model you want (see "Choosing a model" below; 32 GB is comfortable, 16 GB works with small models).
  • LM Studio 0.4.1 or newer, opened at least once.
  • Claude Code: curl -fsSL https://claude.ai/install.sh | bash or brew install --cask claude-code, then claude --version.

Step 1: Load a model in LM Studio

LM Studio ships a command-line tool, lms. If your shell does not find it, run the bootstrap step from the LM Studio app (Developer tab) and reopen the terminal.

lms ls                         # models on disk
lms load --estimate-only qwen/qwen3-coder-30b --context-length 32768
lms load qwen/qwen3-coder-30b --context-length 32768 --identifier qwen-local
lms ps                         # what is loaded, with its context length

Two settings matter more than anything else:

Context length. Claude Code sends a system prompt and tool definitions of roughly 20K tokens before your first word. LM Studio's own guidance is "more than ~25k"; the model's default in LM Studio is often 4K, and with that the session fails on the first request. Use 32K as the floor and 64K if memory allows; the KV cache grows with it.

Tool calling. The agent edits files and runs commands through tool calls. A model without reliable tool use will chat about your code and never touch it. Coding models with tool support that run well on Apple Silicon:

  • qwen/qwen3-coder-30b (30B mixture-of-experts, 3B active; about 19 GB at 4-bit) — the best balance I have found for agentic work.
  • openai/gpt-oss-20b (about 14 GB; LM Studio's and Ollama's default suggestion; needs 16 GB of memory and more with a large context).
  • ibm/granite-4-micro for small machines — usable for simple edits, not more.

Step 2: Start the server

lms server start --port 1234
lms server status

The Anthropic-compatible endpoint is http://localhost:1234/v1/messages. Check it before involving Claude Code:

curl http://localhost:1234/v1/messages \
  -H "content-type: application/json" \
  -H "x-api-key: lmstudio" \
  -d '{"model":"qwen-local","max_tokens":64,
       "messages":[{"role":"user","content":"Say hello in one line."}]}'

A JSON reply with "type":"message" means the server side is done.

Step 3: Point Claude Code at LM Studio

export ANTHROPIC_BASE_URL=http://localhost:1234
export ANTHROPIC_AUTH_TOKEN=lmstudio
export ANTHROPIC_DEFAULT_HAIKU_MODEL=qwen-local
export ANTHROPIC_DEFAULT_SONNET_MODEL=qwen-local
export ANTHROPIC_DEFAULT_OPUS_MODEL=qwen-local
claude --model qwen-local

What each line does:

  • ANTHROPIC_BASE_URL sends every request to LM Studio.
  • ANTHROPIC_AUTH_TOKEN is the credential. LM Studio ignores its value unless Require Authentication is on, in which case use your LM Studio API token. Setting it also tells Claude Code not to use your claude.ai login for the session.
  • The three ANTHROPIC_DEFAULT_*_MODEL lines map Claude Code's model aliases to the local model. This is the step most guides leave out: Claude Code runs some background work on its haiku alias, and if that alias still resolves to an Anthropic model name, LM Studio answers 404 and the session breaks in odd places. With the mapping in place, /model sonnet and friends all mean "the local model".
  • --model qwen-local selects the model for this session; ANTHROPIC_MODEL does the same from the environment.

For a fully offline session add DISABLE_TELEMETRY=1, DISABLE_ERROR_REPORTING=1 and CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1; Claude Code reads only whether these are set, so any value works. LM Studio's integration page also suggests CLAUDE_CODE_ATTRIBUTION_HEADER=0.

Inside Claude Code, /status shows which base URL and credential source are in use — the first thing to check when something is off.

Per-project instead of per-shell

If a project should always use the local model, put the variables in .claude/settings.json in that project (or ~/.claude/settings.json for all projects):

{
  "env": {
    "ANTHROPIC_BASE_URL": "http://localhost:1234",
    "ANTHROPIC_AUTH_TOKEN": "lmstudio",
    "ANTHROPIC_MODEL": "qwen-local",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "qwen-local",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "qwen-local",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "qwen-local"
  }
}

A settings-file value wins over a shell export of the same variable, which is worth remembering when a session ignores what you just exported.

Step 4: Work

cd /path/to/project
claude --model qwen-local

Ask it to inspect the repository, propose edits, run the tests, and review the diff before you commit. Keep tasks small and concrete; local models do far better with "add a --dry-run flag to this script" than with "clean up the project".

Troubleshooting

Connection refused
Nothing listens on the port. lms server status; start the server; confirm the port in ANTHROPIC_BASE_URL.
Claude Code shows the login screen although curl works
The credential variable is not set in the shell you launched claude from, or a settings file overrides it. echo $ANTHROPIC_AUTH_TOKEN; check /status.
A startup warning about two credential sources ("auth may not work as expected")
Both ANTHROPIC_API_KEY and ANTHROPIC_AUTH_TOKEN are set. Keep one; LM Studio accepts either header.
404 or "model not found" in the middle of a session
A background call used the haiku alias. Set ANTHROPIC_DEFAULT_HAIKU_MODEL (and the sonnet/opus ones) to your local model as above.
400 naming a token or context limit
The loaded context is too small. Reload with --context-length 32768 or more; lms ps shows the value in force.
The model explains what it would do but never edits a file
No (working) tool calling. Switch to one of the models listed above.
The first answer takes a minute, later ones are fast
Prompt processing of the ~20K-token system prompt. Normal; larger context and a bigger model make it longer.
Everything is slow
The model does not fit comfortably. Use a smaller quantisation or model, lower the context length, close memory-hungry apps. lms log stream --source server and lms log stream --source model --filter output --stats show what the server is doing.

Summary

  1. lms load <model> --context-length 32768 --identifier qwen-local
  2. lms server start --port 1234
  3. export ANTHROPIC_BASE_URL=http://localhost:1234 ANTHROPIC_AUTH_TOKEN=lmstudio and map the ANTHROPIC_DEFAULT_*_MODEL aliases to the local model
  4. claude --model qwen-local

Related guides: Claude Code with Ollama on macOS and Linux and OpenAI Codex CLI with a local model.

Sources: LM Studio, Claude Code integration; LM Studio, Anthropic compatibility endpoints; Claude Code, connect to an LLM gateway; Claude Code, model configuration.

AItools