---
title: "All Releases"
description: "All notable changes to Nexus OS are documented here. This project follows Semantic Versioning. The foundation release. Everything you need to create, supervise, and scale AI agents from a single CLI. Agent lifecycle mana"
resource: https://www.aiagents.nexus/docs/changelog/releases
generated: { by: "process:nexus-agent-assets", at: 2026-09-07T09:13:03Z }
status: stable
---

# Changelog

All notable changes to Nexus OS are documented here. This project follows [Semantic Versioning](https://semver.org/).

---

## v0.1.0 — Phase 1 Release (2025-04-11)

**The foundation release.** Everything you need to create, supervise, and scale AI agents from a single CLI.

### Core Engine
- Agent lifecycle management: create, start, stop, restart, destroy
- Template system with 5 built-in templates (research, code, data, chat, custom)
- WASM sandbox isolation for agent execution
- SQLite-backed persistent state store
- Project initialization with `naos init`
- YAML-based configuration (`nexus.config.yaml`)

### Orchestration Primitives
- **Supervisors** — one-for-one, one-for-all, rest-for-one restart strategies
- **Sagas** — distributed transactions with compensating rollbacks
- **Workflows** — sequential and parallel step execution with retry policies
- **Agent Pools** — round-robin, least-loaded, and random scheduling

### Cost Control
- Per-agent and global token budgets with hard/soft limits
- Real-time spend tracking across all LLM providers
- Budget alerts at configurable thresholds (50%, 80%, 90%)
- Cost action policies: warn, throttle, pause, kill
- Spending history and analytics

### Trust & AXIS
- AXIS trust verification protocol integration
- Trust tier system: unverified, basic, standard, verified, trusted
- Cryptographic identity verification for agents
- Trust score caching with configurable TTL
- Automatic trust-based access control

### Broker Routing
- Skill-based task routing with pattern matching
- Three-tier routing cascade: Skill → WASM → LLM
- Configurable confidence thresholds (default 0.9)
- Routing statistics and cost savings tracking
- 5 CLI commands: `broker route`, `execute`, `stats`, `config`, `skills`

### Cloudflare Edge Deployment
- Deploy agents as Cloudflare Workers
- Durable Objects for persistent agent state
- Global distribution across 300+ edge locations
- Edge deployment management: deploy, list, status, logs, undeploy
- Simulation mode for testing without live credentials

### Web Dashboard
- 11 server-rendered HTML pages (Overview, Agents, Supervisors, Sagas, Workflows, Pools, Cost, Audit, Trust, Broker, Edge)
- 7 JSON API endpoints for programmatic access
- Terminal/hacker aesthetic with auto-refresh
- `naos dashboard --port 4200 --open`

### Audit Trail
- Append-only audit log for all agent actions
- Timestamped entries with agent ID, action type, and detail
- Queryable via CLI and dashboard

### CLI
- 40+ commands across 8 command groups
- Consistent `--json` output flag for scripting
- Color-coded terminal output with progress indicators
- Global flags: `--help`, `--version`, `--verbose`, `--quiet`

---

## Roadmap

### v0.2.0 — Planned
- Hot-reload agent configuration without restart
- WebSocket real-time dashboard updates
- Multi-provider LLM support (OpenAI, Anthropic, local models)
- Agent-to-agent messaging protocol
- Plugin system for custom orchestration strategies

### v0.3.0 — Planned
- Distributed mode with multi-node clustering
- Kubernetes operator for cloud-native deployment
- Prometheus metrics exporter
- GraphQL API for dashboard
- Visual workflow builder

### v1.0.0 — Planned
- Stable API with backwards compatibility guarantees
- Comprehensive test suite with >90% coverage
- Performance benchmarks and optimization
- Enterprise features: SSO, RBAC, team management
- Commercial support options