Web Console and Agent
rauto has two service entry points that may look similar but serve very different purposes:
rauto webrauto agent
rauto web: local self-managed console
Section titled “rauto web: local self-managed console”rauto web \ --bind 127.0.0.1 \ --port 3000It is designed for single-node use and provides a browser UI for:
- saved connections
- command templates
- command flow templates
- direct command execution
- transaction blocks, workflows, and orchestration
- async task inspection
- command blacklist management
- backup and restore
- recording and replay
Key web console capabilities
Section titled “Key web console capabilities”- connection CRUD and import flows
- SSH compatibility profiles:
secure,balanced, andlegacy-compatible - custom device profiles and prompt diagnostics
- recording, in-browser replay, and execution history
- English and Chinese UI switching
rauto agent: managed execution node
Section titled “rauto agent: managed execution node”rauto agent is not the local UI. It is the managed runtime used to connect an execution node to rauto-manager.
rauto agent \ --bind 0.0.0.0 \ --port 8123 \ --manager-url http://manager:50051 \ --report-mode grpc \ --agent-name agent-beijing-01 \ --agent-token my-secret-token \ --probe-report-interval 300What agent mode provides
Section titled “What agent mode provides”After startup, the agent is responsible for:
- registration and heartbeat with the manager
- inventory reporting and offline notifications
- receiving tasks dispatched by the manager
- sending events and callbacks over HTTP or gRPC
- exposing protected status and probe endpoints
How web and agent differ
Section titled “How web and agent differ”| Command | Audience | Purpose |
|---|---|---|
rauto web | Single-node operator | Local UI, self-managed execution, and resource management |
rauto agent | rauto-manager | A centrally managed execution node |
If you only want to manage templates, run commands, and inspect replays on your own machine, rauto web is enough.
If you want multiple execution nodes under one control plane, use rauto agent.
Agent config file
Section titled “Agent config file”You can also store agent defaults in ~/.rauto/agent.toml:
[manager]url = "http://manager:50051"token = "my-secret-token"report_mode = "grpc"
[agent]name = "agent-beijing-01"heartbeat_interval = 30probe_report_interval = 300This reduces repeated command-line arguments at startup.
Connecting to rauto-manager
Section titled “Connecting to rauto-manager”If the manager uses HTTP reporting:
rauto agent \ --bind 0.0.0.0 \ --port 8123 \ --manager-url http://<manager-host>:3000 \ --report-mode http \ --agent-name edge-sh-01 \ --agent-token <same-agent-api-key>If the manager uses gRPC:
rauto agent \ --bind 0.0.0.0 \ --port 8123 \ --manager-url http://<manager-host>:50051 \ --report-mode grpc \ --agent-name edge-sh-01 \ --agent-token <same-agent-api-key>When to introduce the manager
Section titled “When to introduce the manager”When your requirements start to include the following, it is usually time to move from rauto web to rauto-manager + rauto agent:
- centralized management of multiple execution nodes
- shared inventory and centralized history
- centralized task dispatch
- real-time multi-agent status and task timelines
For more background, continue with rauto-manager Overview.