Roadmap
Completed Features ✅
- Multi-provider support (Google, OpenAI, Anthropic, Ollama, Local)
- Flexible triggers (manual, schedule, watch)
- Built-in tools (file I/O, HTTP)
- Custom tools support
- MCP integration
- File access security (read/write paths)
- Execution tracking and logging
- Schedule triggers (cron + intervals)
- File watching with debouncing
- Hot-reload (auto-reload agent changes)
- Memory & context persistence (session history + long-term memory tools)
- Interactive chat sessions (
agentmd chat) - Skills system (reusable instruction packages, Claude Code compatible)
In Progress 🚧
- Terminal UI (TUI) — Real-time monitoring dashboard
Planned Features 🎯
Near Term (Next 3 months)
- Agent Pipelines — Chain agents together
- Agent Generator — CLI command to scaffold new agents
- Webhook Triggers — HTTP webhooks to trigger agents
- Streaming Output — Stream responses to files/stdout
- Retry Logic — Automatic retry on failures
Medium Term (3-6 months)
- Agent Marketplace — Share and discover agents
- Multi-Agent Conversations — Agents calling other agents
- Vector Memory — RAG-based context retrieval
- Web Dashboard — Browser-based monitoring
- Cloud Deployment — Deploy to cloud platforms easily
Long Term (6+ months)
- Agent Debugging Tools — Step-through execution
- Cost Tracking — Per-agent cost analysis
- A/B Testing — Compare prompt variations
- Agent Observability — Metrics, traces, logs
- Enterprise Features — Teams, permissions, audit logs
Community Requests
Have an idea? Open an issue or start a discussion!
Most Requested
- Agent pipelines / chaining
- Webhook triggers
- Web dashboard
- Cost tracking
How to Contribute
Want to help build these features?
- Check open issues
- Comment on features you'd like to work on
- Submit a PR!
Stay Updated:
Star the GitHub repo
Join Discussions
Report bugs in Issues