O1 Universal Remote is built on a modular, extensible architecture designed to support multiple infrastructure types while maintaining simplicity and power.Documentation Index
Fetch the complete documentation index at: https://o1.network/docs/llms.txt
Use this file to discover all available pages before exploring further.
High-Level Architecture
Core Components
1. API Layer
RESTful HTTP and WebSocket endpoints with token authentication, request validation, and rate limiting.Nodes
Infrastructure lifecycle, SSH keys, cloud credentials — 49 endpoints
Applications
Deployment, versioning, rollback, dependency resolution — 12 endpoints
Monitoring
Metrics, logs, healthchecks, alerting — 22 endpoints
Explorer
Instance pricing, spot history, multi-cloud comparison — 4 endpoints
Addons
Addon CRUD and lifecycle management — 7 endpoints
Operations & Realtime
Playbook execution, PTY sessions, WebSocket streams — 6 endpoints
storage (10), scaling-groups (9), sync (4), deployment (4), environment (4), volumes (1).
2. Service Layer
Nodes Service
Provisions and manages infrastructure nodes across AWS, Azure, GCP, and DigitalOcean. Handles SSH key store, cloud credentials, and infrastructure state sync.
Applications Service
Dual-mode deployment (Docker container or binary). Manages versions, dependencies, configuration templates, and deployment history.
Monitoring Service
Real-time metrics streaming via WebSocket, log aggregation, scheduled healthchecks, and alert dispatch.
Explorer Service
Queries 103K+ indexed instance types for pricing comparison, spot history forecasting, and resource optimization recommendations.
spotScalingService (auto-scaling groups), SyncService (state reconciliation), ptyService (terminal sessions), realtimeService (WebSocket broker).
3. Orchestration Engine
Executes operational playbooks via Ansible (idempotent multi-node automation), Docker (container lifecycle and registry), and Systemd (service management, log rotation, startup ordering). ThePlaybookService resolves dependencies, generates execution plans, tracks state, and writes audit logs.
4. Data Layer
Storage Adapters
Pluggable driver model with four backends: Filesystem (default), MongoDB, S3, and Redis. Configured per-deployment via
StorageConfigService.Configuration
Entities (nodes, applications, addons) persisted as YAML with template variable substitution. Git-trackable by design.
Runtime State
Infrastructure state, deployment history, and monitoring data managed by
infrastructureState and the deployment service.Secrets
SSH keys via
sshKeyStore, cloud credentials via cloudCredentialStore and sharedCredentialStore — never stored in plain config.Data Flow
1. Request Processing
2. Deployment Flow
3. Monitoring Flow
Security Architecture
Authentication
API token auth on all endpoints. SSH keys managed by
sshKeyStore — no passwords stored. Full audit trail on every operation.Secrets & Encryption
Cloud credentials isolated in
cloudCredentialStore. Secrets encrypted at rest and in transit. Network segmentation via firewall rules.Least Privilege
Operations run with minimal required permissions. Automated security hardening applied to provisioned nodes.
Scalability & Performance
Horizontal Scaling
Stateless API services scale independently. External storage backends (MongoDB, Redis, S3) decouple state from compute.
Performance
instanceTypeCache and spotPricing caches reduce upstream API calls. Long-running operations execute asynchronously via playbook queue.High Availability
SyncService reconciles infrastructure state on reconnect. Storage adapter failover supported across all four driver types.Extension Architecture
O1’s adapter pattern makes it straightforward to add new providers, runtimes, and integrations:- Provider adapters — implement the cloud provider interface to add new infrastructure targets
- Storage drivers — extend
StorageDriverto add new persistence backends - Playbook templates — add YAML playbooks to extend the orchestration engine without touching core code
- Webhook events — subscribe to the realtime event bus for external system integration
This modular architecture scales from a single-node homelab to multi-region enterprise deployments — without changing the interface.