Introduction

​

What is ServFlow?

ServFlow is an agent builder. You define AI agents — what they know, which tools they can call, and how the outside world reaches them — and ServFlow serves them from a compiled Go engine. An agent is the main object. Each agent owns:

• Workflows — the units of work it runs, either as inbound HTTP endpoints (webhooks) or as internal and scheduled jobs (tasks)
• A workspace — files the agent can read and write
• Integrations and secrets — the AI models, databases, and APIs it is allowed to use

A workflow is a graph: an entry receives a request, actions do the work, conditionals branch, and responses return a result. The same engine that powers an agent’s tools also serves plain HTTP APIs, so ServFlow works just as well when you only want an endpoint.

​

Why ServFlow?

​

How it works

1. Create an agent — give it a name, and optionally start from a template
2. Add workflows — an HTTP entry for inbound requests, or a trigger entry for internal and scheduled work
3. Add actions — call AI models, query databases, hit external APIs, transform data
4. Configure responses — decide what the workflow returns
5. Run it — locally or on your own infrastructure

​

Three ways to build

ServFlow gives you the same engine through three interfaces — mix them freely: Configurations created in the dashboard are the same objects the CLI reads and writes, so you can prototype visually and then manage the result in code.

​

Deployment options

​

Next steps