Skip to content
This repository was archived by the owner on Jun 28, 2026. It is now read-only.

Commit 9898ced

Browse files
olasunkanmi-SEolasunkanmi-SE
authored
Merge pull request #309 from olasunkanmi-SE/feature_enhancement
update README file
2 parents 9cf0790 + 87a3b9b commit 9898ced

1 file changed

Lines changed: 174 additions & 116 deletions

File tree

‎README.md‎

Lines changed: 174 additions & 116 deletions
Original file line numberDiff line numberDiff line change
@@ -1,144 +1,202 @@
11
# CodeBuddy
22

3-
An AI-powered coding assistant for Visual Studio Code featuring agentic architecture, multi-model support, local model integration, and intelligent codebase understanding.
4-
5-
**CodeBuddy** is an advanced AI-powered VS Code extension featuring a **Multi-Agent System**, **Terminal Integration**, and **Local LLM Support**. It transforms your development workflow by acting as a full-fledged autonomous pair programmer that can plan, execute, test, and debug complex tasks.
6-
7-
## What's New in v3.8.x
8-
9-
- **Terminal Integration** - Agents can now execute shell commands (builds, tests, git) with real-time output streaming and safety guards.
10-
- **Expanded Agent Team** - New specialized agents: **Architect**, **Reviewer**, and **Tester** join the existing team.
11-
- **Collaboration Protocols** - Agents follow structured playbooks for New Features, Bug Fixes, and Refactoring.
12-
- **Local LLM Support** - Full support for Ollama and LM Studio (OpenAI-compatible endpoints).
13-
- **New Model Support** - Added support for Qwen, GLM-4, and DeepSeek-R1/V3.
14-
15-
The extension supports nine AI providers including cloud-based models (Gemini, Anthropic Claude, OpenAI, Groq, DeepSeek, Qwen, GLM) and local models via Ollama, giving developers flexibility in choosing the right model for their privacy and performance requirements.
16-
17-
### The Agent Team
18-
19-
- **Project Manager (Main Agent)**: Orchestrates the workflow and delegates tasks.
20-
- **Architect**: Designs scalable systems, selects patterns, and creates ADRs.
21-
- **Reviewer**: Enforces code quality, security standards, and best practices.
22-
- **Tester**: Writes and executes unit/integration tests to ensure stability.
23-
- **Debugger**: Investigates root causes of errors and proposes fixes.
24-
- **Code Analyzer**: Deeply scans codebase for anti-patterns and complexity.
25-
- **Doc Writer**: Generates comprehensive documentation and API references.
26-
- **File Organizer**: Refactors directory structures and cleans up projects.
27-
28-
### Smart Context Selection
29-
30-
## Core Features
31-
32-
### Terminal Integration (New!)
33-
CodeBuddy can now **run commands directly in your terminal**.
34-
- **Execution**: Run `npm test`, `git status`, `docker build`, or any shell command.
35-
- **Visibility**: Watch real-time output in the dedicated "CodeBuddy Terminal" channel.
36-
- **Safety First**: Every command requires **explicit user confirmation** via a modal dialog before running.
37-
- **Feedback Loop**: Agents read command output (e.g., build errors) to self-correct.
38-
39-
### Collaboration Playbooks
40-
Agents don't just work in isolation; they follow strict protocols:
41-
1. **New Feature**: Architect designs -> Developer implements -> Tester verifies -> Reviewer approves.
42-
2. **Bug Fix**: Debugger finds cause -> Developer fixes -> Tester prevents regression.
43-
3. **Refactoring**: Analyzer identifies issues -> Architect plans -> File Organizer executes -> Tester validates.
44-
45-
### Local LLM Privacy
46-
Run completely offline with local models:
47-
- **Ollama / LM Studio**: Point CodeBuddy to your local server (e.g., `http://localhost:11434/v1`).
48-
- **Privacy**: Your code never leaves your machine when using local models.
49-
- **Zero Cost**: No API fees when running Llama 3, Qwen, or DeepSeek locally.
50-
51-
### Deep Code Understanding
52-
- **RAG & Embeddings**: Vector-based search for semantic understanding of your codebase.
53-
- **Hybrid Storage**:
54-
- `/workspace/`: Real-time file operations.
55-
- `/docs/`: Persistent long-term memory (SQLite).
56-
- `/`: Ephemeral session context.
57-
58-
### Model Context Protocol (Beta)
59-
Connect CodeBuddy to external tools and data sources using the open standard **MCP**.
60-
- **Unified Tooling**: Access database, git, and other tools through a single protocol.
61-
- **Extensibility**: Add any MCP-compliant server to extend CodeBuddy's capabilities.
62-
- **Future-Proof**: Ready for the next generation of AI-tool integration.
3+
CodeBuddy is an AI-powered coding assistant for Visual Studio Code featuring multi-agent architecture, nine AI provider integrations, local model support, and intelligent codebase understanding. It functions as an autonomous pair programmer capable of planning, executing, and debugging complex development tasks.
634

64-
---
5+
## Overview
6+
7+
CodeBuddy enhances developer productivity through AI-powered code assistance, providing intelligent code review, refactoring suggestions, optimization recommendations, and interactive chat capabilities. The extension supports both cloud-based and local AI models, enabling developers to choose the right balance of capability, speed, and privacy for their workflow.
8+
9+
## Key Features
10+
11+
### Multi-Agent System
12+
13+
CodeBuddy employs specialized agents that collaborate on complex tasks:
14+
15+
- **Project Manager**: Orchestrates workflow and delegates tasks to specialized agents
16+
- **Architect**: Designs scalable systems, selects patterns, and creates architecture decision records
17+
- **Reviewer**: Enforces code quality, security standards, and best practices
18+
- **Tester**: Writes and validates unit and integration tests
19+
- **Debugger**: Investigates root causes and proposes fixes
20+
- **Code Analyzer**: Scans codebase for anti-patterns and complexity issues
21+
- **Documentation Writer**: Generates comprehensive documentation and API references
22+
- **File Organizer**: Refactors directory structures and manages project organization
23+
24+
### Dual Operating Modes
25+
26+
**Chat Mode**: Traditional question-and-answer interaction for quick queries, code explanations, and getting code snippets without file modifications.
27+
28+
**Agent Mode**: Autonomous execution with full tool access including file operations, terminal commands, web search, and codebase analysis. Changes can be reviewed before application.
29+
30+
### AI Provider Support
31+
32+
CodeBuddy supports nine AI providers:
33+
34+
| Provider | Default Model | Capabilities |
35+
| --------- | ----------------------- | --------------------------------- |
36+
| Gemini | gemini-2.5-pro | Long context, general coding |
37+
| Anthropic | claude-sonnet-4-5 | Complex architecture, refactoring |
38+
| OpenAI | gpt-4o | Reasoning, planning |
39+
| DeepSeek | deepseek-chat | Cost-effective coding |
40+
| Qwen | qwen-max | Strong open-weight performance |
41+
| Groq | llama-3.1-70b-versatile | Ultra-fast inference |
42+
| GLM | glm-4 | Chinese and English support |
43+
| XGrok | grok | Alternative reasoning |
44+
| Local | qwen2.5-coder | Privacy-first, offline capable |
45+
46+
### Local Model Integration
47+
48+
Run completely offline with local models via Ollama or LM Studio:
49+
50+
- Full Agent mode support with local models
51+
- Zero API costs for unlimited usage
52+
- Code never leaves your machine
53+
- Supports Qwen 2.5 Coder, Llama 3.2, DeepSeek Coder, and other OpenAI-compatible models
54+
55+
### Terminal Integration
56+
57+
Agents can execute shell commands directly:
58+
59+
- Run builds, tests, git operations, and any shell command
60+
- Real-time output streaming in dedicated terminal channel
61+
- Safety guards with user confirmation before execution
62+
- Feedback loop for self-correction based on command output
63+
64+
### Diff Review System
65+
66+
File changes made by the agent are tracked and reviewable:
67+
68+
- Side-by-side diff viewer in VS Code
69+
- Apply or reject individual changes
70+
- Recent changes history panel
71+
- Optional approval mode requiring explicit confirmation before file writes
72+
73+
### Model Context Protocol (MCP)
74+
75+
Extend CodeBuddy's capabilities through the open MCP standard:
76+
77+
- Connect external tools and data sources
78+
- Add custom MCP servers for specialized functionality
79+
- Unified protocol for AI-tool integration
80+
81+
### Project Rules
82+
83+
Customize agent behavior with project-specific rules:
84+
85+
- Define coding conventions and guidelines in `.codebuddy/rules.md`
86+
- Create custom rules via the Settings UI
87+
- Toggle rules on/off without deletion
88+
- Rules are appended to the agent's system prompt
6589

66-
## Quick Start
90+
### Inline Completions
6791

68-
### Installation
69-
You can install **CodeBuddy** from your preferred registry:
70-
- [**VS Code Marketplace**](https://marketplace.visualstudio.com/items?itemName=fiatinnovations.ola-code-buddy)
71-
- [**Open VSX Registry**](https://open-vsx.org/extension/fiatinnovations/ola-code-buddy)
92+
Ghost text suggestions as you type:
7293

73-
Alternatively, search for "CodeBuddy" in your editor's extension manager.
94+
- Configurable completion provider (can differ from chat provider)
95+
- Debounce and token limit controls
96+
- Manual or automatic trigger modes
7497

75-
### Configuration
76-
Configure your preferred AI provider in VS Code Settings:
98+
### Context Management
7799

78-
**1. Cloud Providers (Recommended for best performance)**
79-
- **Gemini**: `gemini-2.5-pro` (Great balance of speed/quality)
80-
- **Anthropic**: `claude-3-5-sonnet` (Top-tier coding capability)
81-
- **Groq**: `llama-3.1-70b` (Insanely fast)
82-
- **DeepSeek**: `deepseek-chat` (Cost-effective)
100+
Intelligent context handling for relevant responses:
101+
102+
- Automatic active file inclusion
103+
- @ mention syntax for adding specific files
104+
- Token budget awareness based on model limits
105+
- Semantic search over indexed codebase
106+
- Priority system: @mentioned files, active file, auto-gathered context
107+
108+
## Installation
109+
110+
Install from your preferred registry:
111+
112+
- [VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=fiatinnovations.ola-code-buddy)
113+
- [Open VSX Registry](https://open-vsx.org/extension/fiatinnovations/ola-code-buddy)
114+
115+
Alternatively, search for "CodeBuddy" in the VS Code extension manager.
116+
117+
## Configuration
118+
119+
### Cloud Providers
120+
121+
1. Open VS Code Settings (File > Preferences > Settings)
122+
2. Search for "CodeBuddy"
123+
3. Select your preferred AI provider from the dropdown
124+
4. Enter your API key in the appropriate field
125+
126+
### Local Models
127+
128+
Configure for Ollama or LM Studio:
83129

84-
**2. Local LLMs (Ollama/LM Studio)**
85-
Set the following in your `settings.json`:
86130
```json
87131
{
88-
"codebuddy.local.baseUrl": "http://localhost:11434/v1", // Default for Ollama
89-
"codebuddy.local.model": "qwen2.5-coder:latest",
90-
"codebuddy.generativeAi.option": "Local"
132+
"generativeAi.option": "Local",
133+
"local.baseUrl": "http://localhost:11434/v1",
134+
"local.model": "qwen2.5-coder"
91135
}
92136
```
93137

94-
---
138+
For Docker-based Ollama:
139+
140+
```
141+
docker compose -f docker-compose.yml up -d
142+
docker exec -it ollama ollama pull qwen2.5-coder
143+
```
95144

96-
## Capabilities & Commands
97-
98-
### Shortcuts
99-
| Command | Shortcut | Description |
100-
| :--- | :--- | :--- |
101-
| **Inline Chat** | `Cmd+I` / `Ctrl+I` | Ask questions about selected code |
102-
| **Review Code** | `Cmd+Shift+R` | Deep quality & security review |
103-
| **Explain** | `Cmd+Shift+E` | Explain complex logic |
104-
| **Add Tests** | `Cmd+Shift+T` | Generate unit tests |
105-
| **Fix Bugs** | `Cmd+Shift+F` | Analyze and fix selected error |
106-
107-
### Supported Models
108-
| Provider | Models | Best For |
109-
| :--- | :--- | :--- |
110-
| **Anthropic** | Claude 3.5 Sonnet, Opus | Complex Architecture, Refactoring |
111-
| **Gemini** | Gemini 1.5 Pro/Flash | Long Context, General Coding |
112-
| **OpenAI** | GPT-4o, o1-preview | Reasoning, Planning |
113-
| **DeepSeek** | DeepSeek-V3, R1 | Cost-Efficiency |
114-
| **Qwen** | Qwen 2.5 Coder | Open Source Leader |
115-
| **Local** | Llama 3, Mistral, etc. | Privacy, Offline |
145+
## Settings Panel
116146

117-
---
147+
Access comprehensive settings via the gear icon in the sidebar:
148+
149+
- **Account**: Profile and account information
150+
- **General**: Theme, language, font settings, nickname, streaming preferences
151+
- **Agents**: Auto-approve actions, file/terminal permissions, verbose logging
152+
- **MCP**: Model Context Protocol server management
153+
- **Conversation**: Chat display preferences and history management
154+
- **Models**: AI model selection and configuration
155+
- **Context**: Workspace indexing, context window size, file inclusion settings
156+
- **Rules and Subagents**: Custom rules and specialized agent configuration
157+
- **Privacy**: Data privacy settings and clear data options
158+
- **Beta**: Experimental features
159+
- **About**: Version information and links
160+
161+
## Data Storage
162+
163+
CodeBuddy stores data locally in a `.codebuddy` folder at your workspace root:
164+
165+
- Chat history in JSON format
166+
- Codebase analysis database (SQLite)
167+
- Project rules and configuration
168+
- Logs and session data
169+
170+
This folder is automatically added to `.gitignore` to prevent committing sensitive data.
118171

119172
## Troubleshooting
120173

121-
**"Command Execution Failed"**
122-
- Ensure you approved the command in the popup modal.
123-
- Check the "CodeBuddy Terminal" output channel for stderr details.
174+
**Connection Issues with Local Models**
124175

125-
**Local LLM Connection Refused**
126-
- Ensure Ollama/LM Studio is running.
127-
- Verify the port (default `11434` for Ollama, `1234` for LM Studio).
128-
- Check `codebuddy.local.baseUrl` setting.
176+
- Verify Ollama or LM Studio is running
177+
- Check the port configuration (default: 11434 for Ollama, 1234 for LM Studio)
178+
- Confirm the `local.baseUrl` setting matches your server
129179

130-
**Agent Loop / Getting Stuck**
131-
- Use the "Stop" button in the chat interface.
132-
- Clear context using the broom icon.
180+
**Agent Not Responding**
133181

134-
---
182+
- Use the Stop button in the chat interface
183+
- Clear chat history via Settings > Privacy
184+
- Check the CodeBuddy output channel (View > Output > CodeBuddy)
185+
186+
**API Key Errors**
187+
188+
- Verify your API key is correctly entered in settings
189+
- Ensure the selected model matches your API key provider
190+
- Check for network connectivity for cloud providers
135191

136192
## Contributing
137-
We welcome contributions! Please see [CONTRIBUTING.md](CONTRIBUTING.md).
193+
194+
Contributions are welcome. Please see [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
138195

139196
## License
140-
MIT License - see [LICENSE](LICENSE).
197+
198+
MIT License - see [LICENSE](LICENSE) for details.
141199

142200
---
143-
**Made with by [Olasunkanmi Raymond](https://olasunkanmi.app)**
144-
_CodeBuddy: Your AI Pair Programmer._
201+
202+
**Repository**: [github.com/olasunkanmi-SE/codebuddy](https://github.com/olasunkanmi-SE/codebuddy)

0 commit comments

Comments
 (0)