Name and switch
The id stays stable for links and history. Turn the scenario off to block real side effects; preview still shows what would happen.
Scenarios
Flows on your board—manual start, clear steps
Scenarios are flows you save on the Digio board: start by hand, call a URL, hand work to an agent, or send a secure callback. Only steps reachable from the start you pick will run. Loops are blocked before anything goes out the door. Preview first, then run for real when you are ready.
Draw on the canvas
Preview first
Pause or stop
Your flow · Run test
Preview and a real test use the same order and log style—preview skips real network calls, new tasks, and webhook posts.
Saved with your team
What we store for each scenario
Each scenario keeps its name, on/off switch, optional default agent, and the map you drew—boxes and the lines between them. Odd steps are gently fixed into a normal start; broken lines are removed so a test never crashes on leftovers.
Default agent
When you set a default specialist, task steps can inherit that person so you do not pick again on every box.
Your canvas map
Each box remembers its place. Lines show which step leads to the next.
Optional schedule
You can attach a schedule when a valid default agent is set—where your workspace supports it.
Building blocks
Five kinds of steps
Each type does one job: start by hand, request the web, queue work for an agent, notify your own HTTPS endpoint, or leave a logo note on the canvas. The editor fills sensible defaults so you move fast.
Manual start
Where a flow begins. If you draw more than one start, you pick which one to test.
Web request
GET or POST with optional JSON body. A live run records status and a short text snapshot of the reply for later steps.
Task for an agent
Queues work with a title, agent, and priority. Preview shows what would be sent; a live run skips quietly if something required is missing.
Webhook notify
POSTs to your HTTPS URL when allowed, with context from the scenario and a short preview of the last web response.
Logo marker
A small brand or integration marker on the canvas—shows in the log for clarity without doing extra work.
Connections
Lines between steps
Lines always go forward: from one step output to another input. Loops and links back into the start are removed. If a port name no longer matches, we fall back to the first valid port so old flows keep working after updates.
Forward from your start
Only steps you can reach from the start you chose will run—handy when you leave experiments on the canvas.
Split into parallel paths
One step can lead to several next steps. When more than one is ready, order is stable and predictable—never a random shuffle.
Broken links
Lines pointing to missing steps are dropped when the map is cleaned up—Run test stays friendly on old drafts.
Order of steps
Reachable steps, sensible order, no loops
We only consider steps connected from your chosen start, then walk forward in a safe order. If the map loops back on itself, we stop and show a clear message instead of spinning forever.
Start not found
The chosen start is missing—often after a bad paste or a deleted box still selected in the list.
No start on the map
Every flow needs at least one manual start before it can run.
Loop detected
A circle in the map cannot finish in order; you will see the issue in the log instead of an endless run.
Web request
Call an API with clear logging
Live runs use a normal web request you can cancel. Answers are read as text and trimmed for the log—no surprise binary dumps.
GET or POST
POST with a body sends JSON when you provide content; otherwise it stays simple.
What you see in the log
Status, success or fail, and size—enough to debug without leaking secrets by default.
Preview line
Preview prints method and URL and marks the step as simulated—great for walkthroughs.
Task step
Give work to a specific agent
We check agent and title before doing anything. Preview echoes what would happen. Live runs match agents on your current roster so stale names do not create mystery tasks.
Priority 1–10
Out-of-range values are brought back into a sensible band.
Runs like pressing "run now"
Tasks created here behave like pressing run on a card—not a silent draft in the backlog.
Long titles allowed
Long titles are preserved up to a generous limit so templates survive import.
Webhook step
Notify your system over HTTPS
When you run for real, we POST compact JSON your server can trust—scenario name, step id, and a short slice of the last web response so you can correlate events.
Allowed URLs
Only secure, allowed URLs are used—the same idea as table status webhooks.
Errors stay visible
Network problems show as readable log lines; cancel stops the run cleanly.
Preview the URL
Preview shows a shortened URL and marks simulation—safe for demos.
Safety
Preview is the default
Opening someone else's scenario starts in preview so webhooks and tasks do not fire by surprise. Disabled flows still explain themselves in the log.
Simulated only
No real request is sent; method and URL appear in the log with a preview label.
No enqueue
Task steps log the would-be title and agent so you can verify fields without touching the board.
No POST
Webhook steps log a safe URL preview and a simulation note.
Still traced
Logo markers still appear in the log so documentation paths stay visible.
Live run
Pause, resume, and cancel
Between steps the runner can wait while you pause or cancel—handy for long requests or demos. Most errors are written to the log so you can recover without losing the whole story.
Checks between steps
Pause and cancel are respected before each piece of work.
Built for tests
The request layer can be swapped in automated tests; in the product you use the browser as usual.
Choosing a start
If no start is selected, the first manual start wins—simple for single-entry flows; pick explicitly when you have branches.
In the overlay
Team, agent filter, integrations
Pick your team, optionally filter flows by default agent, and open integrations when you need credentials—lists stay aligned with the board you are viewing.
Edits before save
Changes stay local until you save; Run test can still use the latest canvas when wired that way.
Open integrations once
A shortcut can expand the credentials panel so you are not hunting menus.
Built-in tips
Hints and shortcuts stay in sync with the rest of the product copy.
Where scenarios plug in
Same agents, tables, and API surface
Task steps use real agents from your roster; web rules match Tables. Heavier scheduled automation also lives in To do, Backlog, and Agent API—Scenarios are great for guided checks you design on the canvas. , Backlog, and Agent API—Scenarios orchestrate manual test hops and cross-system pushes your team designs on the canvas.
Read next
Related product pages
Automation layers, templates with standing tasks, and integrations for inbound triggers.
Rehearse with preview, then go live once
Draw the map, confirm order and skips, then switch to a live run with the same pause habits your team already uses.