Roadmap - djvolz/coda-code-assistant GitHub Wiki
Coda Roadmap
Project Vision
Build a multi-provider, CLI-focused code assistant that provides a unified interface for AI-powered development across Oracle OCI GenAI, Ollama, and other LiteLLM-supported providers.
Current State
✅ Completed Features
Phase 1: Native OCI GenAI Integration
- Support for 30+ models including Cohere Command, Meta Llama, and xAI Grok
- Streaming responses with real-time display
- Dynamic model discovery with 24-hour caching
- Comprehensive authentication methods (API_KEY, SECURITY_TOKEN, INSTANCE_PRINCIPAL, RESOURCE_PRINCIPAL)
Phase 2: Core Provider Architecture
- Abstract provider interface supporting chat, streaming, and model discovery
- LiteLLM integration for 100+ AI providers
- Native Ollama support for local models
- Multi-source configuration system (CLI > env > project > user > defaults)
Phase 3: Enhanced CLI Experience
- Interactive shell with prompt-toolkit
- Slash commands (/help, /model, /provider, /mode, /clear, /exit)
- 7 developer modes (general, code, debug, explain, review, refactor, plan)
- Rich terminal UI with syntax highlighting and tab completion
- Multi-line input support with history
Phase 4: Session Management
- SQLite persistence with automatic migrations (stored in
~/.cache/coda/sessions.db
) - Session commands (/session save/load/list/branch/delete/info/search)
- Export functionality (/export json/markdown/txt/html)
- Full-text search across all sessions
- Context optimization for token limits
- MockProvider for deterministic testing
🚧 Current Development
Phase 5: Tool Integration (MCP) - Target: July 12
- MCP server implementation
- Core tools (file operations, shell, web search, git)
- Tool management commands (/tools)
- Permission management
- Custom tool development SDK
📅 Upcoming Features
Phase 4.5: Automatic Session Saving (Enhancement)
- Auto-save by default with anonymous sessions
- Rolling window management (last 1000 messages)
- User control options and privacy settings
- Performance optimizations with async saves
Phase 6: Advanced Features - Target: July 15
- Multi-modal support (image understanding)
- Document support (PDF, Word, PowerPoint, Excel)
- Enhanced response rendering with live markdown
- Project intelligence with VCS awareness
- UI customization with themes
Phase 7: Web UI (Streamlit)
- Dashboard with provider status and model comparison
- Web-based chat interface
- Usage statistics and cost tracking
- Session management UI
Phase 8: Additional Features
- OpenTelemetry support for monitoring
- Containerization with bundled Ollama
- Repository mapping like Aider
- Automated changelog generation
- Code review helpers
Technical Architecture
Core Dependencies
- Providers: litellm, oci, httpx
- CLI: rich, prompt-toolkit, click
- Storage: sqlalchemy, aiosqlite
- Web: streamlit (optional)
Design Principles
- Async First: Using asyncio throughout for better performance
- Plugin System: Providers and tools as plugins
- Modularity: Self-contained components usable as APIs
- Type Safety: Full type hints and mypy compliance
Configuration
- XDG Base Directory specification
- TOML configuration files
- Environment variable overrides
- Per-project settings support
Release Strategy
Versioning
- Date-based versioning:
year.month.day.HHMM
- Automated releases via conventional commits
- Continuous delivery on main branch
- No manual version management required
Development Workflow
main
: Stable releases (production-ready)develop
: Integration branch (staging)feature/*
: Feature branchesfix/*
: Bug fixes
Recent Milestones
2025.7.5 - Session Management Complete
- Full session persistence with SQLite
- Branching and search capabilities
- Export to multiple formats
- 100+ tests for comprehensive coverage
2025.7.4 - Provider Architecture & Enhanced CLI
- Multi-provider support (OCI, LiteLLM, Ollama)
- Interactive shell with developer modes
- Rich UI features and slash commands
2025.7.3 - Versioning & Release Automation
- Automated GitHub Actions workflow
- Conventional commits support
- Project branding with logos
2025.7.2 - OCI Foundation
- Native OCI GenAI provider
- 30+ model support
- Basic CLI implementation
Contributing
See Development-Guide for setup instructions and contribution guidelines. The project uses conventional commits for automated releases.
Learn More
- Getting-Started - Installation and basic usage
- AI-Modes - Details on the 7 developer modes
- OCI-GenAI-Integration - Oracle Cloud setup guide
- Configuration - Configuration options and examples
- Session-Management - Using sessions and exports