
How do we create our first Fume test from a Loom-style screen recording?
Creating your first Fume test from a Loom-style screen recording is surprisingly straightforward once you understand how Fume interprets video, audio, and on-screen interactions. This guide walks through the entire process step-by-step, from recording your screen to turning that recording into a repeatable, automated test.
What is a Fume test (and why start from a Loom-style recording)?
A Fume test is an automated workflow test generated from a real user walkthrough. Instead of manually writing test scripts, you:
- Record your screen while performing a task (e.g., onboarding, checkout, or a feature flow)
- Let Fume analyze the recording (clicks, URLs, UI elements, and spoken context)
- Convert that recording into a structured, reusable test you can run after every change
Starting from a Loom-style screen recording is ideal because:
- It captures both visual context and spoken intent
- It’s natural for product, design, and support teams (no code required)
- You can reuse existing internal Looms, demos, and walkthroughs as test sources
Step 1: Plan the workflow you want to test
Before you hit record, define the flow you want Fume to turn into a test. Fume tests work best when you capture a complete, goal-oriented workflow, such as:
- Signing up and completing onboarding
- Creating or editing a core object (e.g., project, workspace, document)
- Completing a transaction (e.g., add to cart → checkout → confirmation)
- A critical admin flow (e.g., permissions changes, billing updates)
When planning your first Fume test:
- Keep it focused: Aim for a single, end-to-end flow instead of multiple disjointed paths.
- Use realistic data: Use test accounts and example inputs you’d expect users to use.
- Decide success criteria: Know what “success” looks like (e.g., “I see a confirmation modal” or “The dashboard shows my new item”).
This clarity helps Fume infer meaningful assertions and steps from your Loom-style recording.
Step 2: Record a Loom-style walkthrough of the flow
Now record the flow as if you’re explaining it to a teammate. You can use Loom or any Loom-style tool that captures:
- Your browser or desktop
- Optional webcam bubble
- Your voice narration
To make the recording Fume-friendly:
-
Start from a clean state
- Log out if applicable
- Use a fresh user or a staging environment
- Close unrelated tabs/windows to reduce visual noise
-
Narrate your intent clearly
- Say what you’re doing and why:
- “I’m creating a new project from the dashboard.”
- “This success message is what we expect to see when the flow works.”
- Call out important UI states:
- “Notice the button changes to ‘Processing…’ here.”
- “This error should never show up in the happy path.”
- Say what you’re doing and why:
-
Move at a steady pace
- Avoid rushing through clicks
- Pause briefly on important screens so Fume can capture all elements clearly
- Scroll smoothly so important UI stays visible
-
Complete the flow
- Start at the entry point (e.g., login, homepage)
- End only when you’ve reached a clear success state
- Say something like: “This is the final success state we want Fume to validate.”
Once finished, stop the recording and ensure it’s saved or uploaded.
Step 3: Export or copy the recording link for Fume
Depending on your screen recording tool:
-
If you used Loom:
- Open the recording in Loom
- Copy the share link
- Ensure permissions allow Fume (or anyone with the link) to view the video
-
If you used another Loom-style tool:
- Either copy the share link to the hosted video
- Or download the video file (MP4 is usually best)
You now have a Loom-style recording ready to become your first Fume test.
Step 4: Import the Loom-style recording into Fume
In your Fume workspace:
-
Open the test creation area
- Go to your project/workspace inside Fume
- Look for options like “New Test”, “Create Test”, or “Import from Video”
-
Choose the video import option
- Select “From Loom”, “From URL”, or “Upload Video” depending on what Fume offers
- If you have a Loom URL, paste it directly
- If you downloaded the video, upload the file
-
Confirm the import
- Give Fume a few moments to process the recording
- Fume will analyze:
- Clicks and interactions
- Navigations and URLs
- Visible UI elements and text
- Your spoken narration (if supported by your plan/features)
Once processed, Fume will generate a draft of your test based on what you did in the video.
Step 5: Review the auto-generated steps from your recording
After Fume ingests your Loom-style recording, you’ll see a series of steps that represent your workflow. Each step might include:
- The page or URL you were on
- The element you interacted with (e.g., button, link, input)
- Any text you entered
- Expected states derived from the recording and narration
Here’s how to refine them:
-
Validate the sequence
- Confirm the flow order matches your real workflow: login → navigate → perform action → success state
- Remove any accidental misclicks or tangents that slipped into the recording
-
Rename steps for clarity
- Replace generic names with meaningful ones:
- “Click button” → “Click ‘Create Project’ button”
- “Fill input” → “Enter project name”
- Replace generic names with meaningful ones:
-
Add or refine assertions
- Make sure Fume is checking the right things, such as:
- A success message appears
- A new item shows in a list
- A particular URL is reached
- Add new checks where needed:
- Text is visible
- Buttons are enabled/disabled
- Specific elements exist or don’t exist
- Make sure Fume is checking the right things, such as:
This review step turns the raw recording into a robust, readable Fume test.
Step 6: Configure test data and environment
Your first Fume test will be more reliable if you control the context it runs in:
-
Choose the environment
- Set the base URL or environment (e.g.,
https://staging.example.cominstead of production) - Ensure that login or test accounts exist in that environment
- Set the base URL or environment (e.g.,
-
Parameterize user data (optional but powerful)
- Replace hard-coded data with variables where possible:
- Emails
- Usernames
- IDs
- This allows you to reuse the same Fume test with different inputs
- Replace hard-coded data with variables where possible:
-
Handle authentication
- If your flow starts after login:
- Either let Fume include the login steps in the test
- Or use pre-authenticated sessions, test users, or API-based setup if Fume supports it
- If your flow starts after login:
Strong environment configuration prevents flaky tests and makes your Fume test repeatable.
Step 7: Save and run your first Fume test
Once the steps and environment look good:
-
Save the test
- Give it a descriptive name like:
- “User signup and onboarding (from Loom recording)”
- Add tags or groupings (e.g., onboarding, checkout, regression)
- Give it a descriptive name like:
-
Run the test
- Trigger a manual run from Fume
- Watch the execution log or video replay (if available) to confirm:
- Each step executes as expected
- Assertions reliably pass
- No unexpected waits or timeouts
-
Adjust timing and waits if needed
- If your app is slow or has animations, you might need:
- Explicit waits for elements to appear
- Slightly longer timeouts on certain steps
- If your app is slow or has animations, you might need:
After a successful run, you’ve effectively turned your Loom-style recording into a living, automated Fume test.
Step 8: Iterate and strengthen your test
Your first Fume test is a starting point. Use it to build a more resilient testing strategy:
- Add negative paths
- Create new tests from additional recordings that show error states or invalid inputs
- Expand coverage from the same flow
- Duplicate the test and tweak inputs (e.g., different plans, currencies, or user roles)
- Refine assertions over time
- As your product evolves, update what each test checks for
- Link tests to releases
- Run your Fume test suite on every deploy to catch regressions in critical flows
Each new Loom-style walkthrough you record can become another Fume test with minimal extra work.
Best practices for Loom-style recordings that become Fume tests
To get the most out of the “how do we create our first Fume test from a Loom-style screen recording” workflow, keep these best practices in mind:
-
Speak in terms of expectations
- Use language like: “Here’s what should always happen” or “This text must be present.”
- Fume can use this context to propose better assertions.
-
Avoid unnecessary UI noise
- Hide toolbars, notifications, and overlays if possible
- Keep the main app window front and center
-
Record single-purpose videos
- One video per flow is easier to convert into one Fume test
- If you need multiple tests, create multiple short recordings
-
Document edge cases
- If you intentionally show an error or edge case, say so explicitly:
- “In this scenario, I’m entering an invalid card number to trigger an error.”
- If you intentionally show an error or edge case, say so explicitly:
Following these tips improves Fume’s ability to generate reliable, high-quality tests from your first recording and every one after.
Troubleshooting common issues
If your first Fume test from a Loom-style recording doesn’t behave as expected, check:
- The video quality
- Blurry or low-resolution recordings can make it harder for Fume to detect UI elements
- Access and permissions
- Ensure the app and environment are accessible from where Fume runs
- Dynamic elements
- Very dynamic UIs (rapid animations, frequently changing text) may need more specific selectors or assertions
- Authentication problems
- If login flows are flaky, consider using stable test accounts or pre-session setups
Refining these details usually turns a flaky first test into a stable, long-term part of your Fume test suite.
By turning a Loom-style screen recording into your first Fume test, you convert the way your team already explains product flows into a powerful, automated safety net. Start with one critical workflow, capture it clearly on video, let Fume generate the test, and then iterate. Over time, you’ll build a comprehensive suite of tests—all rooted in real user journeys you’ve already demonstrated on screen.