> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.meetgail.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.meetgail.com/_mcp/server.

# Nodes & Connections

> How individual steps and the arrows between them combine into a workflow.

A workflow is a set of steps joined by arrows. Each step, called a node, does
one job: send an email, look up a contact, decide which way to go. The arrows,
called connections, decide the order those steps run in. Together they turn a
list of separate actions into one automated process that runs on its own.

## Nodes are the steps

Every node is a single building block with a clear purpose. Some nodes talk to
outside services, some transform data, and some make decisions. You add a node,
give it the settings it needs, and connect it to the rest of your workflow.

A few nodes have special roles:

* **Start** is where every run begins. It receives the information a run is
  given and hands it to the first step. See
  [Starting a Workflow](/platform/workflows/core-concepts/starting-a-workflow).
* **End** marks where a path finishes. A workflow can have more than one End if
  different paths wrap up in different ways.

Everything in between is up to you. Browse the full set in the
[Node Reference](/platform/workflows/node-reference).

## Connections set the order

A connection is an arrow from one node to the next. It says "when this step
finishes, do that step next." Because the arrows define the order, they also
define what data is available at each point: a step can only use values produced
by steps that run before it. See
[Variables & Data Flow](/platform/workflows/core-concepts/variables-data-flow) for how
that works.

A simple workflow is often just a straight line:

```
Start -> Look up contact -> Send email -> End
```

## Branching, looping, and grouping

Not every workflow is a straight line. A few nodes change the shape of the run:

* **[Conditional](/platform/workflows/node-reference/logic-and-flow/conditional)** is a
  fork in the road. It sends the run down different paths depending on the data,
  so high-value accounts and routine ones can be handled differently.
* **[For Each](/platform/workflows/node-reference/logic-and-flow/for-each)** repeats a set
  of steps once for every item in a list, such as every contact in an uploaded
  file.
* **[Sub Graph](/platform/workflows/node-reference/logic-and-flow/sub-graph)** tucks a
  whole workflow inside a single node, so you can reuse a sequence of steps and
  keep a large workflow readable.

Give each node a clear name that says what it does, like "Score applicant" or
"Notify agent." Names show up in run history and in the value picker, so good
names make a workflow much easier to follow later.

## Related

#### [Variables & Data Flow](/platform/workflows/core-concepts/variables-data-flow)

How values move along the connections between steps.

#### [Node Reference](/platform/workflows/node-reference)

Browse every node, grouped by what it does.