Bugzi - AI-Powered Multi-Agent Coding Companion for VS Code
Bugzi is a powerful, multi-agent AI assistant that integrates directly into your VS Code workflow. Unlike traditional code assistants, Bugzi can handle complex development tasks step-by-step — creating and editing files, exploring your entire project, running terminal commands, and more. With built-in real-time security scanning, intelligent code completion, and support for the latest AI models from every major provider, Bugzi is your all-in-one development companion.
A free alternative to Cursor AI, GitHub Copilot, Windsurf, Bolt.new, v0.dev, Replit Agent, Cline, Codeium, Augment Code, Cody, Continue.dev, Aider, PearAI, Tabnine & Amazon CodeWhisperer.
Features
|
Bugzi can use your CLI and Editor to accomplish complex tasks autonomously:
- Create & Edit Files: Write new code or modify existing files with a rich diff view. Monitor linter/compiler errors along the way to proactively fix issues like missing imports and syntax errors.
- Execute Terminal Commands: Run shell commands directly in your VS Code terminal and monitor their output. Perfect for installing packages, running builds, executing tests, or deploying applications.
- File Operations: Read files, write files, make targeted SEARCH/REPLACE edits, list directories with interactive tree navigation, and search across files using regex patterns — all with rich, color-coded response cards.
- Workspace Exploration: Automatically analyze file structure, source code ASTs, and run regex searches to understand your project context without overwhelming the context window.
- Real-Time Error Detection: Pre-process AI results to detect and fix malformed code blocks early, reducing error-prone output.
|
Screenshot: AI creating and editing files with diff view
|
Real-Time Security Scanner
Screenshot: Built-in OWASP/CWE/CVE scanning
|
Real-Time Security Scanning
Built-in security scanner that identifies and patches vulnerabilities as you code:
- OWASP/CWE/CVE Detection: Automatically scans your code for known security loopholes and vulnerable patterns.
- SQL Injection & XSS Prevention: Catches injection vulnerabilities before they make it to production.
- Language Server Integration: Java-based JAR language server for deep code analysis and linting.
|
Multi-Model AI Support
Bugzi supports every major AI provider and lets you switch between models on the fly:
| Provider |
Models |
| Anthropic |
Claude Fable, Opus 5/4.8/4.7/4.6/4.5/4.1/4, Sonnet 5/4.6/4.5/4, Haiku 4.5/3.5 |
| OpenAI |
GPT-5.6/5.5/5.4/5.3, GPT-5.2/5.1 Codex, o4-mini, o3-pro/o3/o3-mini, GPT-4.1, GPT-4o |
| Google |
Gemini 3.7/3.6/3.5 Flash, Gemini 3.5 Flash Lite, Gemini 3.1 Pro, Gemini 2.5 Pro/Flash/Lite, Gemma 4/3 |
| DeepSeek |
DeepSeek V4 Pro (incl. 0813), DeepSeek V4 Flash (incl. vision), DeepSeek Chat |
| Qwen (Alibaba) |
Qwen3.8/3.7/3.6/3.5, Qwen3 Coder/Next, Qwen3 VL, Qwen 2.5 |
| Mistral AI |
Mistral Medium 3.5, Codestral, Devstral, Mistral Large/Small, Ministral |
| xAI |
Grok 4.6/4.5/4.3/4.20, Grok Build |
| MiniMax |
MiniMax M3, M2.x, M1, MiniMax-01 |
| Moonshot (Kimi) |
Kimi K3, K2.7, K2.6, K2.5, K2 |
| Z.ai (GLM) |
GLM 5.3/5.2/5.1/5, GLM-4.7/4.6v |
| + more |
ByteDance Seed (Doubao), NousResearch |
Cloud: Amazon Bedrock, Google Cloud Vertex AI
Local: Ollama, LM Studio, LiteLLM, VS Code LM API
Router: OpenRouter (auto-fetches latest models)
|
AI Model Selector
Screenshot: Selecting from 200+ AI models across providers
|
|
Screenshot: Color-coded file operation cards with syntax highlighting
|
- Color-Coded Tool Cards: Every file operation renders as a distinct, color-coded card with expand/collapse sections, status badges, and per-operation metadata.
- Tree-sitter Syntax Highlighting: Accurate highlighting across 20+ languages — automatically adapts to your VS Code theme.
- AI Loading Animations: 12 distinct animation styles provide visual feedback during AI processing.
- Interactive Directory Tree: Browse your project structure with an expand/collapse tree component directly in chat.
- Settings Panel: Clean UI for configuring autocomplete, model selection, and provider settings.
|
Context & Memory Management
- Context Configuration Dialog: Fine-tune AI model parameters and project context for each conversation.
- Chat History: Browse, search, and revisit past conversations with a dedicated history page.
- Checkpoint System: Automatic workspace snapshots at each step — compare diffs and restore to any previous state with confidence.
- Token & Cost Tracking: Monitor total tokens and API usage cost for the entire task loop and individual requests.
|
Chat History & Context
Screenshot: Browsing chat history and managing context
|
Inline Code Completion
Screenshot: AI autocomplete suggesting code as you type
|
Inline Code Completion (Ghost Text)
- Context-aware autocomplete that suggests code as you type.
- Configurable debounce delay and maximum completion lines.
- Choose between Bugzi Cloud or Local (Ollama) provider.
- Code Lens Integration: "Bugzi: Complete Code" button above your cursor to manually trigger AI code generation.
|
MCP Server Support (Experimental Features)
Use the Model Context Protocol to extend Bugzi's capabilities with custom tools:
- Connect community-made MCP servers.
- Ask Bugzi to create and install custom tools tailored to your workflow.
- Supports file system, Git, GitHub, Notion, Atlassian, Figma, Azure, and more.
"Add a tool that fetches Jira tickets" — and Bugzi handles everything from creating the server to installing it.
|
MCP Server Configuration
Screenshot: Custom tool integration via MCP
|
Getting Started
Installation
- Open Visual Studio Code.
- Go to the Extensions view (
Ctrl+Shift+X).
- Search for "Bugzi".
- Click Install.
- The Bugzi icon will appear in your activity bar.
Quick Start
- Click the Bugzi icon in the activity bar to open the chat panel.
- Select your preferred AI model from the dropdown.
- Type your task — for example:
- "Create a REST API with Express and TypeScript"
- "Find and fix bugs in this function"
- "Add a tool that fetches Jira tickets"
- "Scan this file for security vulnerabilities"
- Review and approve file changes and terminal commands as Bugzi works through your task.
- Use Checkpoints (Compare/Restore) to track changes at every step.
Context Commands
Use @ to give Bugzi additional context:
@url — Paste a URL to fetch and convert to markdown (great for docs).@file — Add a file's contents directly to context (type to search files).@folder — Add an entire folder's contents at once.@problems — Add workspace errors/warnings for Bugzi to fix.
Configuration
| Setting |
Description |
Default |
bugzi.autocomplete.enabled |
Enable/disable AI autocomplete |
true |
bugzi.autocomplete.debounceDelay |
Delay (ms) before triggering autocomplete |
300 |
bugzi.autocomplete.maxLines |
Max lines for autocomplete suggestions |
3 |
bugzi.autocomplete.provider |
Autocomplete provider (Bugzi Cloud or Local) |
Bugzi Cloud |
bugzi.autocomplete.localModel |
Model name for Local/Ollama provider |
qwen2.5-coder:1.5b |
Security
Bugzi is designed with a human-in-the-loop approach:
- Every file change is presented as a diff view for your review before applying.
- All terminal commands require your explicit approval before execution.
- Built-in security scanning identifies vulnerabilities as you code.
- All changes are recorded in your file's Timeline for easy tracking and rollback.
Contributing
We welcome contributions! Here's how to get started:
- Fork the repository.
- Read our Contributing Guide (coming soon).
- Join our Discord to chat with other contributors.
- Submit a pull request with your changes.
License
This project is licensed under the Apache 2.0 License.
Support & Feedback
Enterprise
Get the same Bugzi experience with enterprise-grade controls:
- SSO (SAML/OIDC) authentication
- Global policies and configuration
- Observability with audit trails
- Private networking (VPC/private link)
- Self-hosted or on-premises deployments
- Enterprise support
Learn more about Bugzi Enterprise (coming soon)
| |
