<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/i-built-claude-code-s-workflow-with-a-local-model-and-it-s-surprisingly-close-tlmlhtfbc" -->

---
title: I built Claude Code's workflow with a local model, and...
description: A developer built a DIY alternative to Claude Code using LM Studio to run Qwen 3.5 (14B, 4-bit quantized) locally on an M5 MacBook Pro with 24GB unified...
canonical: https://daily.dev/posts/i-built-claude-code-s-workflow-with-a-local-model-and-it-s-surprisingly-close-tlmlhtfbc
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: I built Claude Code's workflow with a local model, and it's surprisingly close | daily.dev
og:description: A developer built a DIY alternative to Claude Code using LM Studio to run Qwen 3.5 (14B, 4-bit quantized) locally on an M5 MacBook Pro with 24GB unified...
og:url: https://daily.dev/posts/i-built-claude-code-s-workflow-with-a-local-model-and-it-s-surprisingly-close-tlmlhtfbc
og:image: https://api.daily.dev/og/posts/tlmlHTFBC.png
og:image:alt: I built Claude Code's workflow with a local model, and it's surprisingly close
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# I built Claude Code's workflow with a local model, and it's surprisingly close

**[XDA Developers](https://daily.dev/sources/xda-developers)** · 3 min read · 0 upvotes · 0 comments

## Summary

A developer built a DIY alternative to Claude Code using LM Studio to run Qwen 3.5 (14B, 4-bit quantized) locally on an M5 MacBook Pro with 24GB unified memory, paired with Hermes as the agentic coding tool. The setup splits Claude Code's single-product workflow into three pieces: Qwen for decision-making, LM Studio for model serving, and Hermes for executing file edits, terminal commands, and test runs. Testing on a focused configuration bug showed the local agent could search code, edit files, run tests, and self-correct failing tests without manual error copy-pasting, though each step was slower than Claude Code, especially with larger files or terminal output. The author notes local agents have improved significantly and are viable on consumer hardware like a MacBook Air with 16GB+ RAM, or larger 32B models with a dedicated GPU and enough VRAM.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.xda-developers.com/built-claude-code-workflow-local-model-surprisingly-close>

## Questions this post answers

### Can I run a local model to replicate Claude Code's agentic coding workflow for free?

Yes, by combining LM Studio to serve a local model through an OpenAI-compatible API with an agent tool like Hermes to execute file edits, terminal commands, and tests. One setup ran Qwen 3.5's 14B model (4-bit quantized, 32,768 token context) on an M5 MacBook Pro with 24GB unified memory, handling focused coding tasks like fixing a configuration bug and self-correcting a failing test.

_daily.dev surfaces practical local AI agent setups for developers comparing free alternatives to Claude Code._

### What hardware do I need to run a good local coding model for free?

A MacBook Pro or even a MacBook Air with more than 16GB of RAM can run a capable local model, and a machine with a dedicated GPU and enough VRAM can handle a 32-billion-parameter model. A 14B Qwen 3.5 model quantized to 4-bit ran comfortably on an M5 MacBook Pro with 24GB of unified memory.

_developers weighing local versus cloud AI coding tools can track hardware guidance like this on daily.dev._

## Similar posts on daily.dev

- [I ran Qwen3.5 locally instead of Claude Code. Here’s what happened.](https://daily.dev/posts/i-ran-qwen3-5-locally-instead-of-claude-code-here-s-what-happened--fnmwj9ey5) · InfoWorld · 0 upvotes · 0 comments
- [Claude Code with a local LLM running offline is the hybrid setup I didn't know I needed](https://daily.dev/posts/claude-code-with-a-local-llm-running-offline-is-the-hybrid-setup-i-didn-t-know-i-needed-nj1yzrce2) · XDA Developers · 0 upvotes · 0 comments

---

Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#claude-code](https://daily.dev/tags/claude-code), [#local-ai](https://daily.dev/tags/local-ai), [#qwen](https://daily.dev/tags/qwen), [#lm-studio](https://daily.dev/tags/lm-studio)

[View this post on daily.dev](https://daily.dev/posts/i-built-claude-code-s-workflow-with-a-local-model-and-it-s-surprisingly-close-tlmlhtfbc)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"I built Claude Code's workflow with a local model, and it's surprisingly close","url":"https://daily.dev/posts/i-built-claude-code-s-workflow-with-a-local-model-and-it-s-surprisingly-close-tlmlhtfbc","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/i-built-claude-code-s-workflow-with-a-local-model-and-it-s-surprisingly-close-tlmlhtfbc"},"datePublished":"2026-09-09T21:31:24.399Z","dateModified":"2026-09-09T21:32:24.838Z","description":"A developer built a DIY alternative to Claude Code using LM Studio to run Qwen 3.5 (14B, 4-bit quantized) locally on an M5 MacBook Pro with 24GB unified...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/26476ab59cfef4304f109a090edef7e9?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/26476ab59cfef4304f109a090edef7e9?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"XDA Developers","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"XDA Developers","logo":"https://media.daily.dev/image/upload/s--kCg6yyAP--/f_auto,q_auto/v1774964407/logos/xda-developers?_a=BAMAMiWQ0","url":"https://daily.dev/sources/xda-developers"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/i-built-claude-code-s-workflow-with-a-local-model-and-it-s-surprisingly-close-tlmlhtfbc","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,claude-code,local-ai,qwen,lm-studio","timeRequired":"PT3M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"XDA Developers","item":"https://daily.dev/sources/xda-developers"},{"@type":"ListItem","position":3,"name":"I built Claude Code's workflow with a local model, and it's surprisingly close"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/i-built-claude-code-s-workflow-with-a-local-model-and-it-s-surprisingly-close-tlmlhtfbc#faq","mainEntity":[{"@type":"Question","name":"Can I run a local model to replicate Claude Code's agentic coding workflow for free?","acceptedAnswer":{"@type":"Answer","text":"Yes, by combining LM Studio to serve a local model through an OpenAI-compatible API with an agent tool like Hermes to execute file edits, terminal commands, and tests. One setup ran Qwen 3.5's 14B model (4-bit quantized, 32,768 token context) on an M5 MacBook Pro with 24GB unified memory, handling focused coding tasks like fixing a configuration bug and self-correcting a failing test. daily.dev surfaces practical local AI agent setups for developers comparing free alternatives to Claude Code."}},{"@type":"Question","name":"What hardware do I need to run a good local coding model for free?","acceptedAnswer":{"@type":"Answer","text":"A MacBook Pro or even a MacBook Air with more than 16GB of RAM can run a capable local model, and a machine with a dedicated GPU and enough VRAM can handle a 32-billion-parameter model. A 14B Qwen 3.5 model quantized to 4-bit ran comfortably on an M5 MacBook Pro with 24GB of unified memory. developers weighing local versus cloud AI coding tools can track hardware guidance like this on daily.dev."}}]}
```

