The latest articles on DEV Community by Ryan (@breakingdawm). https://dev.to/breakingdawm https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F1208092%2F665256a4-9642-4d9b-a209-106bf5264eb6.png https://dev.to/breakingdawm en Ryan Tue, 05 May 2026 15:52:16 +0000 https://dev.to/breakingdawm/grom-free-open-source-ai-coding-assistant-for-vs-code-ollama-lm-studio-anthropic-and-more-4m1l https://dev.to/breakingdawm/grom-free-open-source-ai-coding-assistant-for-vs-code-ollama-lm-studio-anthropic-and-more-4m1l <h1> Grom — Free, Open-Source AI Coding Assistant for VS Code (Ollama, LM Studio, Anthropic, and More) </h1> <p>I've been building <strong>Grom</strong>, a free and open-source VS Code extension that brings agentic AI coding to your machine. No telemetry, no mandatory account, no subscription. If you use Ollama or LM Studio, nothing ever leaves your machine.</p> <h2> What is it? </h2> <p>Grom is a chat + agentic coding extension that lives in the VS Code sidebar. You can talk to it like a regular AI assistant, or switch it into <strong>BUILD mode</strong> where it reads files, writes code, searches your codebase, and runs terminal commands autonomously — pausing for your approval before anything destructive happens.</p> <h2> Key features </h2> <p><strong>Chat</strong></p> <ul> <li>Streaming chat with <strong>PLAN</strong> and <strong>BUILD</strong> modes</li> <li>Multiple sessions, compact history, export to Markdown</li> <li>Persistent memory — custom instructions injected into every chat</li> <li>Per-session system prompt override</li> </ul> <p><strong>Agentic loop</strong></p> <ul> <li>Built-in file tools: read, write, delete, search, list directory, run terminal</li> <li>MCP (Model Context Protocol) server support</li> <li>Per-action approval for destructive operations — nothing gets written or deleted without you seeing it first</li> <li>Diff-aware undo — after the agent writes files, an undo button appears so you can revert any or all of the changes</li> <li>Task log showing every tool call with args and results</li> </ul> <p><strong>RAG</strong></p> <ul> <li>Your codebase is automatically indexed with BM25 + optional semantic embeddings via Ollama</li> <li>Relevant files are attached to every message without you having to think about it</li> </ul> <p><strong>Inline autocomplete</strong></p> <ul> <li>Ghost-text completions as you type</li> <li>Adaptive debounce — slows down automatically when you rarely accept suggestions</li> <li>Word-by-word partial accept</li> <li>Per-language model routing</li> </ul> <p><strong>@ Context mentions</strong></p> <ul> <li> <code>@filename</code> — attach any workspace file</li> <li> <code>@selection</code> — currently selected code in the editor</li> <li> <code>@git</code> — your uncommitted diff</li> <li> <code>@terminal</code> — recent terminal output</li> <li> <code>@problems</code> — all VS Code errors and warnings</li> <li> <code>@url:https://...</code> — fetch and attach a web page</li> <li> <code>@docs</code> — search indexed documentation sources</li> </ul> <p><strong>Providers</strong><br> Works with pretty much everything:</p> <ul> <li> <strong>Local</strong>: Ollama, LM Studio, Open Code</li> <li> <strong>Cloud</strong>: Anthropic (Claude), OpenAI (GPT-4o), Groq, Mistral, Gemini</li> <li> <strong>Custom</strong>: any OpenAI-compatible or Anthropic-compatible endpoint</li> </ul> <p>API keys are stored in the OS keychain — never in <code>settings.json</code>.</p> <h2> Why I built it </h2> <p>I wanted something like Cursor or GitHub Copilot but with full control over where my code goes. Most existing extensions either require a cloud subscription, send your code to a third party by default, or don't support local models well. Grom is designed so that local-first is the default, and cloud providers are opt-in extras.</p> <h2> Quick start </h2> <ol> <li>Install from the <a href="https://marketplace.visualstudio.com/items?itemName=RyanConnolly.grom" rel="noopener noreferrer">VS Code Marketplace</a> </li> <li>Install <a href="https://ollama.com" rel="noopener noreferrer">Ollama</a> and run <code>ollama pull qwen2.5-coder</code> </li> <li>Open the Grom panel from the activity bar and start chatting</li> </ol> <p>For cloud providers, select one from the dropdown and paste your API key when prompted — it gets stored in the OS keychain immediately.</p> <h2> Links </h2> <ul> <li> <strong>Marketplace</strong>: <a href="https://marketplace.visualstudio.com/items?itemName=RyanConnolly.grom" rel="noopener noreferrer">https://marketplace.visualstudio.com/items?itemName=RyanConnolly.grom</a> </li> <li> <strong>GitHub</strong>: <a href="https://github.com/ryanjames85/grom" rel="noopener noreferrer">https://github.com/ryanjames85/grom</a> </li> <li> <strong>Docs</strong>: <a href="https://ryanjames85.github.io/grom" rel="noopener noreferrer">https://ryanjames85.github.io/grom</a> </li> </ul> <p>It's early days — v0.3.5 is out now. Would love feedback, bug reports, or feature requests via GitHub Discussions.</p> ai opensource showdev vscode