LocalMinds Logo
New: Build your own AI Workflows — Agents, Skills & Loops

Code Smarter.
Stay Private.

LocalMinds brings powerful AI-assisted coding to VS Code. Run fully offline with local models or tap into 200+ cloud models — your choice, your control.

Download Free Explore Features
200+
Cloud Models
100%
Private Offline
Free
To Get Started
localminds — VS Code Terminal
$ localminds chat
✓ Connected to local model (llama3.2:3b) — offline mode active
You: Refactor this function to use async/await
LocalMinds
Sure! Here's the refactored version using async/await:
async function
fetchUserData(id) {
const data = await api.get(`/users/${'{id}'}`);
return data.json();
}
█Your code never left your machine
New · Headline Feature

Build Your Own AI Workflows

Define reusable, multi-step AI tasks that live in .localminds/ alongside your code and are committed with your repo — so your whole team benefits from the same automations.

.localminds/agents/implementer.yaml
name: Implementer
model: anthropic/claude-sonnet-4
mode: edit

system: |
  You implement features precisely
  following the approved plan.

Chain blocks into a Workflow

Ordered steps where each can reference an earlier step's output via {{steps.ID.output}}. Repeat a fixed number of times, or loop until a condition is met.

.localminds/workflows/ship-feature.yaml 4 steps · 1 loop
1
Agent

Plan the change

gpt-4o-mini

Reads the linked issue and drafts a step-by-step implementation plan.

2
Skill

Write failing tests

Invokes the "Test Author" skill using {{steps.1.output}} to cover new behaviour.

3
Agent

Implement & fix

claude-sonnet-4

Edits files to satisfy the tests, re-running them after every change.

Loop back to step 2 until tests pass (max 5)
4
Instruction

Lint & format check

Applies repo conventions automatically before the diff preview is shown.

Watch it run

Live streaming output per step, a diff preview you approve before any file changes, and a Cancel button to stop at any point.

Run outputStep 3 of 4
Plan the change — complete
Write failing tests — complete
Implement & fix — streaming…
Lint & format check
Diff preview
src/api/users.ts+3 −1
− return fetch(url).then(r => r.json())
+ const res = await fetch(url)
+ if (!res.ok) throw new Error('Request failed')
+ return res.json()

Review every change before it touches your files.

Run LocalMinds: Project Workflows from the Command Palette

First run seeds a small example workflow in .localminds/ at your project root to get you started.

⌘⇧P → Project Workflows
New

Also new this release

Workflows is the headline, but it shipped alongside a handful of smaller upgrades to how you plan, see, and steer the AI.

Plan Mode

A dedicated, read-only chat mode for structured thinking before writing code — concrete step-by-step plans for refactors and architecture decisions, no file edits applied.

66% used

Context Window Indicator

A ring gauge beside the input box breaks usage down into history, system prompt, and files — hover for exact token counts.

Redesigned Mode Selector

Toggle buttons are now a dropdown that shows a short description next to each mode — the difference between Ask, Plan, and Agent is always visible.

Ask (was “Chat”)PlanAgent

Per-Step Model Override

Run a cheap, fast model for boilerplate steps and a more capable one only where it matters. Each step cancels independently, without interrupting a parallel chat stream.

Features

Built for Developer Control

Everything you need. Nothing you don't. No lock-in, no surprises.

Complete Privacy

Your code never leaves your machine when using local models. Zero telemetry, full data ownership.

100% offline capable

Lightning Fast

Local models deliver instant responses with zero network latency. Blazing fast completions, anywhere.

<50ms response time

Ultimate Flexibility

Switch between local models or 200+ cloud providers on the fly. Full cost control, no vendor lock-in.

200+ model providers

AI Code Generation

Generate, refactor, explain, and debug code using natural language. Works with all major languages.

Multi-language support

Cost Control

Use free local models to eliminate API costs entirely, or choose cloud models that fit your budget.

$0 with local models

VS Code Native

Deep VS Code integration with context-aware suggestions, inline diffs, and side panel chat.

Deep IDE integration
How It Works

Up and running in minutes

1

Install the Extension

Search for LocalMinds in the VS Code Extensions marketplace and click Install.

2

Choose Your Model

Set up Ollama for free offline models, or add your OpenRouter API key for 200+ cloud options.

3

Start Coding with AI

Open the side panel, highlight code, or use inline commands. LocalMinds is ready instantly.

Model Support

Your Models, Your Rules

Start local for free, scale to cloud when needed

Local Models

via Ollama — 100% offline

Free
llama3.2:3b
llama3.2:latest
codellama:7b
deepseek-coder:6.7b
mistral:7b
phi3:mini

+ many more via Ollama library

Cloud Models

via OpenRouter — 200+ models

Flexible
OpenAI GPT-4ogpt-4o, gpt-4o-mini
Anthropic Claudeclaude-3.5-sonnet
Google Geminigemini-2.0-flash
Meta LLaMAllama-3.3-70b
DeepSeekdeepseek-r1, v3

Pay only for what you use

Get Started

Install in seconds

Available on the VS Code Marketplace — free to install.

Install via VS Code CLI

ext install localminds.localminds
or

Search "LocalMinds" in VS Code Extensions panel

Open VS Code Marketplace
LocalMindsLocalMinds·Privacy-First AI

© 2026 LocalMinds. Built for developers who care about privacy.