Connect agents

A connection is a directed edge from one agent to another. It tells the orchestrator: "after the source finishes, the target reads its output and runs next."

Drawing a connection

Hover over an agent. Two small handles appear on its left and right edges. Click and drag from the right edge of the source agent to the left edge of the target. A connection line draws between them.

Connections are always left-to-right; they imply ordering. If you try to draw a connection going backward, Cerevisor blocks it. (Loops are a separate connection type, see Control flow.)

Smart auto-connect

Cerevisor automatically wires agents in adjacent columns. This is auto-connect, and it pairs agents 1-to-1 by vertical position: the topmost agent in a column connects to the topmost agent in the next column, the second to the second, and so on. If one column has more agents than the next, the extra agents connect to (or from) the last agent in the shorter column.

Auto-connections have type auto; connections you draw by hand have type manual. If you draw a manual connection for a pair that auto-connect would have made, your manual wire wins and auto-connect skips that pair. When auto-connections are recalculated, only auto connections are touched; manual wires are always preserved.

You can exclude specific source-to-target pairs from auto-connect via the workflow's autoConnectionExclusions setting in Settings → Workflow.

Handoff formats

Click a connection (or double-click) to open the Connection Settings popup. The Handoff Format field appears only when the workflow has structured orchestration / native mode enabled:

Format	What downstream sees
Summary	A concise summary of the upstream agent's output. Useful when the full output is long and you only need the gist.
Structured JSON	The upstream output as a JSON object with a defined shape. Use when the downstream agent needs specific fields.
File Reference	The upstream agent's output files — by name, not content. The downstream agent reads them via the read_file tool. Useful when output is large; keeps context lean.
Full Output	The upstream agent's complete output, verbatim (clipped at roughly 4k tokens per connection — avoid on agents with many upstream feeds).

If you don't pick one, Cerevisor chooses automatically: File Reference when the upstream agent declares output files, otherwise Summary.

The right format depends on context budget. Long workflows with chatty agents benefit from Summary or File Reference; pick Full Output only when downstream truly needs everything.

Data description

A free-text field on the connection. Describes what data flows over this edge. The text is included in the generated prompt so downstream agents understand the intent. Example: "Vetted candidate sources with confidence scores."

You don't have to fill it. Cerevisor generates a sensible default from the source agent's output definition.

Selecting and editing connections

• Click an edge to select it.
• Shift-click to multi-select.
• Double-click to open Connection Config.
• Right-click for the connection context menu (Configure / Flow Type / Delete).
• Delete / Backspace with edges selected removes them.

Connection patterns

Auto-connect produces 1-to-1 positional pairings between adjacent columns. To build the other common shapes, draw manual wires:

1. One-to-many fan-out: one upstream agent feeds N downstream agents in the next column. All N start in parallel.
2. Many-to-one fan-in: N upstream agents feed one downstream. The downstream waits for all N to finish, then runs.
3. All-to-all: N upstream, M downstream. Each downstream gets all N upstream outputs.

Manual wires also let you carve exceptions to the auto-connect pairing (e.g. "Reviewer A only reads from Researcher 1, not Researcher 2").

Excluding pairs from auto-connect

For a workflow where each column should produce siloed outputs that the next column picks from selectively, exclude the relevant source-to-target pairs from auto-connect in Settings → Workflow. Then draw manual wires.

Connection colors

Connections inherit their color from the source agent's role accent. A connection from a Developer (emerald) → Reviewer (violet) draws an emerald-to-violet gradient. The color tells you at a glance which role originated the data flowing over the edge.