OpenAgents
Getting Started
Quick Start Guide SDK Reference
Framework
Psionic Web Framework Architecture Overview
Contributing
Development Guide Roadmap
Support
Troubleshooting
⚠
Documentation is a work in progress

This documentation is in very early stages and actively being developed. Content may be incomplete or subject to change.

Roadmap

Current state and future plans for OpenAgents

Roadmap

This document outlines the current state of OpenAgents and our vision for the future. As an early-stage project, priorities may shift based on community feedback and technical discoveries.

Important: OpenAgents is heavily in development. APIs and features are subject to change.

Current State (June 2025)

✅ Implemented

Core SDK

  • Basic agent creation: Identity generation with placeholder Nostr keys
  • Ollama integration: Local AI inference with streaming support
  • TypeScript-first: Full type safety and modern tooling
  • Effect.js foundation: Service architecture ready for expansion

Web Framework

  • Psionic: Server-side rendering with component explorer
  • WebTUI: Terminal-inspired UI components with theming
  • Documentation site: Built with our own tools

AI Capabilities

  • Local inference: Privacy-preserving AI through Ollama
  • Streaming responses: Real-time token generation
  • Embeddings: Semantic search and similarity
  • Multi-model support: Any Ollama-compatible model

Developer Experience

  • Monorepo structure: Clean package boundaries
  • Hot reload: Fast development cycles
  • TypeScript: End-to-end type safety
  • Documentation: Comprehensive guides and API reference

🚧 In Development

Nostr Integration

  • NIP-06 key derivation: Deterministic identities from mnemonics
  • Event handling: Type-safe Nostr event system
  • Relay connections: WebSocket management
  • Status: Nostr package exists but not integrated with SDK

AI Providers

  • Claude integration: For MAX subscribers via CLI
  • Provider abstraction: Unified interface
  • Status: AI package implemented, needs SDK integration

📋 Placeholder/Stub Features

Economic Model

  • Lightning payments: Invoice generation returns stub data
  • Agent economics: Lifecycle states defined but not implemented
  • Bitcoin integration: Vision established, implementation pending

Infrastructure

  • Container deployment: Firecracker VM integration planned
  • Compute management: Resource allocation stubs
  • Distributed agents: Multi-node deployment concepts

Short Term (Q3 2025)

1. Complete Core Integration

  • [ ] Wire Nostr package into SDK for real key generation
  • [ ] Integrate AI package for multiple providers
  • [ ] Implement basic agent persistence
  • [ ] Add agent communication via Nostr

2. Developer Experience

  • [ ] Improve error messages and debugging
  • [ ] Add more examples and tutorials
  • [ ] Create project templates
  • [ ] Enhance component explorer

3. Testing & Quality

  • [ ] Increase test coverage to 80%+
  • [ ] Add integration test suite
  • [ ] Implement visual regression tests
  • [ ] Performance benchmarks

4. Documentation

  • [ ] API documentation generator
  • [ ] Video tutorials
  • [ ] Architecture deep dives
  • [ ] Migration guides

Medium Term (Q4 2025 - Q1 2026)

1. Lightning Network Integration

  • [ ] Real Lightning invoice generation
  • [ ] Payment flow implementation
  • [ ] Wallet integration guides
  • [ ] Micropayment streaming

2. Agent Capabilities

  • [ ] Tool use (function calling)
  • [ ] Memory and context management
  • [ ] Multi-agent coordination
  • [ ] Task scheduling

3. Infrastructure

  • [ ] Container isolation (Firecracker)
  • [ ] Resource monitoring
  • [ ] Auto-scaling agents
  • [ ] Distributed deployment

4. Ecosystem

  • [ ] Plugin system for agent extensions
  • [ ] Marketplace for agent templates
  • [ ] Community agent registry
  • [ ] Integration templates

Long Term Vision (2026+)

1. Autonomous Agent Economy

  • Self-sustaining agents: Earn Bitcoin to cover operational costs
  • Market dynamics: Supply and demand for agent services
  • Reputation system: Trust and quality metrics
  • Agent evolution: Natural selection based on profitability

2. Technical Innovation

  • Zero-knowledge proofs: Privacy-preserving agent interactions
  • Homomorphic encryption: Compute on encrypted data
  • Decentralized compute: P2P resource sharing
  • Cross-chain integration: Beyond Bitcoin/Lightning

3. Use Cases

  • Personal assistants: AI agents you truly own
  • Business automation: Self-managing service providers
  • Research agents: Autonomous scientific investigation
  • Creative collaborators: AI partners for content creation

4. Standards & Protocols

  • Agent communication protocol: Standardized message formats
  • Capability discovery: How agents find each other
  • Payment negotiation: Automated pricing mechanisms
  • Identity verification: Trust without centralization

How to Contribute

Immediate Needs

  1. Testing: Write tests for existing functionality
  2. Documentation: Improve guides and examples
  3. Bug fixes: Check GitHub issues
  4. Examples: Build demo applications

Feature Development

  1. Lightning integration: Help implement real payments
  2. Nostr features: Enhance protocol support
  3. AI providers: Add new inference backends
  4. UI components: Expand WebTUI library

Research Areas

  1. Agent economics: Model sustainable pricing
  2. Distributed systems: Multi-agent coordination
  3. Security: Adversarial agent scenarios
  4. Performance: Optimize resource usage

Design Principles

As we build toward this vision, we maintain these principles:

  1. Local-first: User control over data and compute
  2. Open protocols: No vendor lock-in
  3. Economic alignment: Agents must provide real value
  4. Privacy preserving: Minimal data exposure
  5. Developer friendly: Great tooling and documentation

Get Involved

Community

Development

Feedback

  • Feature requests: Open a GitHub issue
  • Bug reports: Include reproduction steps
  • Documentation: PRs always welcome
  • Ideas: Start a discussion

Disclaimer

This roadmap represents our current vision and is subject to change. Timelines are estimates, not commitments. We prioritize based on:

  1. Community feedback
  2. Technical feasibility
  3. Resource availability
  4. Strategic alignment

The best way to influence priorities is to contribute code, documentation, or constructive feedback.

Building the future of autonomous AI agents, one commit at a time. 🚀