Concepts

Sinfonia separates what runs (the bot), when it runs (the trigger) and where it runs (the agent). The orchestrator, in the cloud, stores the bots and coordinates executions. Agents, installed on your machines, run the code.

    flowchart LR
    G["Trigger<br/>manual · schedule · files · AWS"] -->|fires| E["Execution<br/>bot + version + parameters"]
    E --> F[("Queue")]
    F --> A["Agent<br/>on your machine"]
    A -->|logs and status| O["Monitoring<br/>dashboard · executions"]

You publish a bot (via Script, Zip or Git) and Sinfonia keeps every version.
A trigger decides when the bot should run and creates an execution.
The execution enters the queue until the selected agent is available.
The agent runs the bot on its machine and streams logs and status in real time.
You follow everything on the dashboard and the executions screen.

Glossary

Term	What it is
Bot	Your automation: the code and dependencies published to Sinfonia. It can be Python, JavaScript (Node), binary (EXE/ELF), Shell, Batch or PowerShell.
Version	Each bot publication gets an `X.Y.Z` version. In triggers you pick a fixed version or `release`, which always uses the latest one.
Agent	Program installed on a Windows or Linux machine that runs the bots. It connects to the orchestrator over HTTPS.
Trigger	Rule that fires a bot execution: manual, scheduling, files and folders or AWS.
Execution	One run of a bot, with status (`SUCCESS`, `ERROR`, `RUNNING`, `STOPPED`), logs and duration.
Queue	Executions waiting for an available agent.
Parameters	JSON set on the bot or trigger. It reaches the code as environment variables.
Environment Variables	Values and secrets registered once in the account and used by many bots.
Mode	Whether the bot runs without a graphical interface (`Headless`) or with one (`Headed`).

Where to go next

Your First Bot

See the full cycle in practice in a few minutes

Choose a trigger

Which trigger to use in each situation

Last updated on 27/09/2026