Supported Agents
Compatibility matrix and feature comparison for AI coding agents supported by RAPID.
Agent Comparison
Section titled “Agent Comparison”| Feature | Claude Code | OpenCode | Aider | Copilot CLI |
|---|---|---|---|---|
| Provider | Anthropic | Multi | Multi | GitHub |
| Interface | Chat CLI | Chat CLI | Chat CLI | Commands |
| File Editing | Yes | Yes | Yes | No |
| Git Integration | Yes | Yes | Yes | No |
| Multi-file | Yes | Yes | Yes | No |
| Auto-commit | No | No | Yes | N/A |
| MCP Support | Yes | Yes | No | No |
| Context Files | CLAUDE.md | AGENTS.md | .aider.conf.yml | N/A |
Detailed Agent Information
Section titled “Detailed Agent Information”Claude Code
Section titled “Claude Code”Developer: Anthropic
License: Commercial
Models: Claude 3.5 Sonnet, Claude 3 Opus
Strengths
Section titled “Strengths”- Excellent reasoning capabilities
- Strong at complex refactoring
- Good understanding of context
- MCP server integration
Best For
Section titled “Best For”- Architecture decisions
- Complex code changes
- Code review
- Documentation
Configuration
Section titled “Configuration”{ "claude": { "cli": "claude", "instructionFile": "CLAUDE.md", "envVars": ["ANTHROPIC_API_KEY"], "installCmd": "npm install -g @anthropic-ai/claude-code" }}Pricing
Section titled “Pricing”- Based on API token usage
- See Anthropic pricing
OpenCode
Section titled “OpenCode”Developer: Community
License: MIT
Models: Claude, GPT-4, GPT-3.5, Local models
Strengths
Section titled “Strengths”- Multi-provider support
- Extensible architecture
- Open source
- MCP integration
Best For
Section titled “Best For”- Flexibility in model choice
- Cost optimization
- Custom workflows
Configuration
Section titled “Configuration”{ "opencode": { "cli": "opencode", "instructionFile": "AGENTS.md", "envVars": ["ANTHROPIC_API_KEY", "OPENAI_API_KEY"], "installCmd": "npm install -g opencode" }}Pricing
Section titled “Pricing”- Based on underlying provider pricing
- Can use local models for free
Developer: Paul Gauthier
License: Apache 2.0
Models: GPT-4, GPT-3.5, Claude, Local models
Strengths
Section titled “Strengths”- Automatic git commits
- Strong pair programming workflow
- Efficient token usage
- Voice mode support
Best For
Section titled “Best For”- Quick code changes
- Iterative development
- Auto-committing workflow
Configuration
Section titled “Configuration”{ "aider": { "cli": "aider", "instructionFile": ".aider.conf.yml", "envVars": ["OPENAI_API_KEY"], "installCmd": "pip install aider-chat", "args": ["--model", "gpt-4o"] }}Model Options
Section titled “Model Options”{ "args": ["--model", "gpt-4o"] // OpenAI GPT-4}{ "args": ["--model", "claude-3-5-sonnet-20241022"] // Anthropic}Pricing
Section titled “Pricing”- Based on underlying provider pricing
- Token-efficient architecture
GitHub Copilot CLI
Section titled “GitHub Copilot CLI”Developer: GitHub/Microsoft
License: Commercial (GitHub subscription)
Models: OpenAI Codex
Strengths
Section titled “Strengths”- GitHub integration
- Explain commands
- Generate shell commands
Limitations
Section titled “Limitations”- No file editing
- No multi-file context
- Command-focused only
Best For
Section titled “Best For”- Shell command generation
- Git operations
- Quick explanations
Configuration
Section titled “Configuration”{ "copilot": { "cli": "gh", "args": ["copilot"], "envVars": ["GITHUB_TOKEN"], "installCmd": "gh extension install github/gh-copilot" }}Pricing
Section titled “Pricing”- Included with GitHub Copilot subscription
Model Comparison
Section titled “Model Comparison”| Model | Provider | Context | Speed | Cost | Best For |
|---|---|---|---|---|---|
| Claude 3.5 Sonnet | Anthropic | 200K | Fast | $$ | General coding |
| Claude 3 Opus | Anthropic | 200K | Slow | $$$ | Complex reasoning |
| GPT-4o | OpenAI | 128K | Fast | $$ | Quick tasks |
| GPT-4 Turbo | OpenAI | 128K | Medium | $$ | Balanced |
| GPT-3.5 Turbo | OpenAI | 16K | Fast | $ | Simple tasks |
Feature Details
Section titled “Feature Details”File Editing
Section titled “File Editing”How agents modify files:
| Agent | Method | Confirmation |
|---|---|---|
| Claude Code | Direct write | Shows diff first |
| OpenCode | Direct write | Configurable |
| Aider | Direct write | Optional confirm |
| Copilot CLI | N/A | N/A |
Git Integration
Section titled “Git Integration”| Agent | Auto-stage | Auto-commit | Custom messages |
|---|---|---|---|
| Claude Code | Yes | No | N/A |
| OpenCode | Yes | No | N/A |
| Aider | Yes | Yes | Yes |
| Copilot CLI | No | No | No |
Context Window Usage
Section titled “Context Window Usage”How agents use available context:
| Agent | Strategy |
|---|---|
| Claude Code | Full context + instruction file |
| OpenCode | Selective context loading |
| Aider | Repository map + focused files |
| Copilot CLI | Current command only |
Choosing an Agent
Section titled “Choosing an Agent”Use Case Recommendations
Section titled “Use Case Recommendations”| Use Case | Recommended Agent |
|---|---|
| New feature development | Claude Code |
| Bug fixing | Aider |
| Code review | Claude Code |
| Quick refactoring | Aider |
| Architecture decisions | Claude Code |
| Cost-sensitive work | OpenCode (GPT-3.5) |
| Local/offline | OpenCode (Ollama) |
| Git workflow | Aider |
Adding New Agents
Section titled “Adding New Agents”RAPID can work with any CLI-based AI tool:
{ "agents": { "available": { "custom-agent": { "cli": "my-ai-tool", "instructionFile": "MY_AGENT.md", "envVars": ["MY_API_KEY"], "installCmd": "npm install -g my-ai-tool", "args": ["--flag", "value"] } } }}Requirements
Section titled “Requirements”For an agent to work with RAPID:
- CLI interface - Must be runnable from command line
- Environment variables - Must read API keys from env
- Interactive mode - Must support chat/interactive session
- Exit cleanly - Must handle SIGTERM gracefully
Troubleshooting
Section titled “Troubleshooting”Agent not starting
Section titled “Agent not starting”# Verify installationwhich claudewhich aider
# Check env varsecho $ANTHROPIC_API_KEY
# Reinstallrapid start --reinstall-toolsSlow response times
Section titled “Slow response times”- Check model selection (Opus slower than Sonnet)
- Reduce context window
- Use faster model for simple tasks
Rate limiting
Section titled “Rate limiting”- Implement request queuing
- Use multiple API keys
- Switch to different provider
Future Agents
Section titled “Future Agents”Planned support:
| Agent | Status | ETA |
|---|---|---|
| Cursor CLI | Planned | TBD |
| Continue | Evaluating | TBD |
| Cody | Evaluating | TBD |
| Local LLMs (Ollama) | Via OpenCode | Available |