The AI engineering platform that already knows what tools you have.

ProjectMind AI automatically discovers your AI environment, learns from your codebase over time, and brings multi-agent code review, intelligent task routing, cost tracking, and autonomous code editing directly into VS Code.

What it does

When you open a project with ProjectMind:

Scans your environment — detects your API keys, local Ollama models, MCP servers, and other AI tools automatically. No manual configuration.
Builds a capability matrix — decides which model is best for coding, security review, documentation, reasoning, etc. based on what you actually have.
Shows a live sidebar — Brain, Git Intelligence, Cost tracking, and Execution Plans update as you work.
Runs multi-agent reviews — 3 specialized agents (Architect, Security, Quality) review code in parallel and synthesize a single verdict.

Features

Panel	What it shows
Repository Brain	Which files are hotspots, tech debt by category, contributor quality scores
Git Intelligence	Commit type breakdown (feat/fix/refactor/docs), file churn, recent PR risk scores
Cost & Budget	Live spend tracking, monthly forecast, budget alerts, model tier breakdown
Execution Plans	AI-generated multi-step plans waiting for your approval

Commands

Open with Cmd+Shift+P and type "ProjectMind":

Command	Description
Discover AI Environment	Scans your machine and shows which models, providers, and agents were detected. Builds an intelligent routing table automatically.
Import Detected Providers	One-click import of all detected API keys into the encrypted provider store
Multi-Agent Review	Runs Architect + Security + Quality agents in parallel on any diff
Generate Agent Prompt	Generates a context-enriched prompt for any task using your project's full context
Score PR Risk	6-factor risk assessment on a diff: size, file types, test coverage, security, complexity, history
Approve Plan	Review and approve AI-generated execution plans step by step
Set Monthly Budget	Configure a spend limit — the platform automatically downgrades models to stay within budget

Getting started

1. Install the extension

Search "ProjectMind AI" in the VS Code Extensions panel, or install from the command line:

code --install-extension aaravkhanal.projectmind

2. Start the backend

The extension connects to the ProjectMind backend running locally. Clone the repo and start it:

git clone https://github.com/Aaravkhanal/llm-reviewer
cd llm-reviewer
.venv/bin/uvicorn backend.api.main:app --host 0.0.0.0 --port 8000 --reload

Or with Docker:

docker compose up

3. Open a project

Open any project folder in VS Code. Click the ProjectMind brain icon in the Activity Bar — the sidebar loads automatically.

4. Discover your environment

Press Cmd+Shift+P → ProjectMind: Discover AI Environment

The extension scans your machine and shows:

Which editor/IDE you're running
Which API providers were detected (from environment variables)
Which local models are running (Ollama, LM Studio, etc.)
Which MCP servers are configured
The routing table: which model handles each task type

Settings

Setting	Default	Description
`projectmind.backendUrl`	`http://localhost:8000`	URL of the ProjectMind backend
`projectmind.llmProvider`	`nvidia`	Default LLM provider for reviews
`projectmind.reviewBudgetUsd`	`1.0`	Max spend per review in USD