To record a stream, screen or gameplay session with Streamlabs Desktop, prepare the sources you want, check the recording options under Settings > Output, and press Record in the editor. You do not need to start a YouTube broadcast to make a local recording.
When you stop, Streamlabs says it shows a completion notification once the file is saved. Check that notice and open the file before relying on the recording; a button press alone does not tell you whether the picture and sound came out as intended.
Choose recording during initial setup
When you first configure Streamlabs Desktop, choose the setup option that describes recording videos if that is your main purpose. Recording is presented alongside live streaming, so selecting it helps guide the initial setup towards the job you want to do. You can record without going live.
The exact screens and labels can change between versions. If you have already completed setup, there is no need to start over: you can configure recording later in Settings > Output and work from the editor. Streamlabs' getting-started guide documents its current onboarding flow; compare its instructions with the version installed on your computer.
It helps to decide what the recording is for before choosing sources. A full-screen tutorial needs the display or application you are demonstrating, while narration needs a microphone source. A session you plan to edit may need microphone and computer sound kept separate; a simple review copy may be easier as one mixed recording. Those goals affect the checks you make before starting, not whether you need to begin a live broadcast.
If the eventual goal is an always-on channel rather than a saved file, recording in Streamlabs Desktop and broadcasting continuously are different tasks. A local recording depends on the desktop session and chosen output path. For a prerecorded product demonstration that you intend to broadcast, this guide to streaming product setup tutorials covers a different workflow.
Add the source you need to capture
Choose a capture source according to scope. Display Capture is for the whole monitor. Window Capture is for a particular application. If you choose a window, open the target application and keep it unminimised while selecting it; a closed or minimised window may not appear in the selection list.
Before recording anything important, look at the editor preview. Confirm that it shows the intended display or application, and that notifications, unrelated windows or private information are not visible. If you are recording a tutorial, open the pages or files you will demonstrate and arrange them first. This small preparation is easier than trimming mistakes out later.
Some applications may require permissions, or may prevent their contents from being captured. Do not infer success from the source name appearing in the list. Check the preview and make a short test recording, then open that file and confirm the actual image. Streamlabs' PC screen-recording guide describes display and window capture; restrictions can depend on the application and your system.
For narration, add or select a microphone source. Your computer's built-in microphone may be adequate for a basic voice-over; an external or USB microphone is optional, not a prerequisite. If you want your voice and computer sound, add the relevant audio sources where supported and check their levels. Avoid adding the same source twice, since duplicated audio can produce echo or a doubled signal.
A webcam is optional. Add it only if the recording needs it, then check its framing in the preview. A clear desktop capture with no webcam is often the better choice for a screen demonstration. The goal is to include only the sources that serve the recording, rather than filling the scene with inputs you have not checked.
Configure recording options in Settings > Output
Open Settings > Output and review the recording destination and the available filename, quality and format options. Set a destination you can find again and that you can write to. If you choose a removable drive or a folder with restricted permissions, confirm that it is available and writable before a longer take.
The settings shown depend on the installed version and system. Do not assume there is one universal preset for every computer or recording purpose. Streamlabs' recording tutorial and settings discussion offers starting points, but its recommendations are not guarantees of quality or performance on your machine.
Think through the trade-offs rather than copying a preset without a test. Higher output resolution or quality can make a recording more demanding to produce and store. Frame rate is another choice: fast gameplay has different motion needs from a mostly static lesson or devotional visual. Choose a sensible starting point for the content and hardware you have, then inspect the result and adjust only if needed.
Streamlabs' tutorial discusses matching the base canvas to monitor resolution for gameplay, considering 60 FPS for fast action or 30 FPS for more static material, and using Lanczos as a quality-oriented downscale filter with additional CPU cost. These are examples from the vendor, not a promise that they suit every setup. If the computer struggles, reduce output resolution or adjust quality and test again. Do not treat any of these numbers as a guaranteed result.
Encoding choices also involve hardware and workload. The tutorial describes x264 as CPU-based software encoding and NVENC as a hardware option for NVIDIA graphics cards. It discusses CQP as a starting rate-control method, CRF with x264, a CQ starting point of 17 and the Max Quality preset where performance allows. Treat these as options to test, not universal ideals; Streamlabs itself advises checking system performance as quality is increased.
| Decision | What to consider | Practical check |
|---|---|---|
| Capture scope | Whole display or one application | Confirm the preview contains the intended area |
| Resolution and frame rate | Detail and motion versus the computer's workload | Record a short sample and review playback |
| Encoding option | CPU use or supported graphics hardware | Check that the test recording completes smoothly |
| Audio arrangement | One mixed sound or separate tracks for editing | Play the file and confirm each needed source is present |
| File format | Simplicity or a workflow that needs multitrack audio | Check the selected format against the editing requirement |
If you need separate microphone and desktop audio for editing, use the documented multitrack workflow rather than assuming the default recording separates inputs. Streamlabs instructs users to change Output mode to Advanced, select Recording, choose audio tracks, and assign inputs in the advanced audio mixer. Its guide documents up to six separate tracks. It also says MP4 or MKV is needed for multitrack audio because the default FLV format does not support multiple audio tracks. Review the official multitrack instructions for the steps and current interface.
Separate tracks are useful if you expect to adjust narration independently in editing, but they make setup and file review more important. If you are recording and streaming at the same time, Streamlabs notes that streaming sends one selected audio track. Check that the track used for the live mix still contains the sound you intend viewers to hear. If you only need a straightforward local copy, a multitrack arrangement may add complexity without helping.
Prepare the scene and check audio
Before the main take, arrange the scene and make a short test. Confirm the right source is visible, the cursor is where you expect it, and any presentation material is ready. If you are recording a long lesson or loop, check the beginning and a later point in the content. A source may look right in the preview but behave differently when you play the saved recording.
Watch the audio indicators while speaking or playing sound. Make sure the microphone is responding if you need narration, and that computer sound is present if the recording requires it. Then play the test file, not only the live preview. This catches issues such as an inactive source, an echo from duplicate inputs or a silent application capture.
On macOS, system-audio capture has a particular limitation in Streamlabs' documented setup: macOS does not allow Streamlabs Desktop to directly capture audio from other applications. Streamlabs describes routing system sound through a virtual audio device such as BlackHole and selecting it as a Mic/Auxiliary device. Consult its Mac audio guide for current steps and check compatibility with your macOS version and audio-routing software. Do not assume the Windows-style desktop-audio approach applies unchanged.
A brief test is also a useful way to judge whether your chosen quality settings suit the computer. If playback stutters or the recording does not complete smoothly, change one relevant option at a time and test again. That makes it easier to identify a workable balance than changing resolution, frame rate and encoding settings together.
Press Record in the editor
Once the preview and test are satisfactory, use Record in the editor to begin. Streamlabs describes recording as a control separate from going live, so pressing it starts a local recording rather than automatically starting a YouTube broadcast. Confirm that the recording has actually begun before proceeding with a presentation or session.
For a tutorial, pause briefly after starting before speaking or switching screens. That gives you a clean opening to trim if needed. During a longer take, keep an eye on the scene and audio indicators when practical. If you change sources mid-recording, verify the new image and sound rather than assuming the previous checks still apply.
A recording session is only as reliable as the computer and storage available to it. Avoid starting a long take when the selected destination is nearly full or when you are not sure you can access it. This is particularly worth checking before an overnight capture: use a short test and make sure you can open the result before leaving the setup unattended.
If you are capturing material for an always-on YouTube channel, a local file may be a useful source asset, but it is not itself a 24/7 broadcast. The comparison of a VPS and a spare PC for a continuous stream explains the separate operating decision. For a channel that needs to keep a prerecorded file live while your own computer is switched off, StreamNeo removes that specific need to leave the desktop recording or playback setup running: you upload the video and use your YouTube stream key to begin a YouTube-only broadcast.
Stop and locate the saved recording
When you have finished, press the recording control again to stop. Streamlabs says a completion notification appears when the file has been saved. Use that notification to inspect or open the recording rather than relying only on the fact that the control changed state.
The selected recording destination in Settings > Output is the first place to check if you need to locate the file later. If you set a filename option, use that to distinguish this take from other recordings. The exact folder and file name are choices you make in the output options; there is no single location that applies to every installation.
If you cannot find the file, return to the configured destination and check the name and folder you chose. Confirm that the destination is connected and accessible. A notification or open-file action is a more direct confirmation than guessing from a default path that may not match your setup.
Keep the working file somewhere you can identify and access for any later edit or upload. If the recording matters, make a copy using your usual file-management process after you have confirmed the original plays. That does not change how Streamlabs records; it simply reduces the chance that a later move or cleanup leaves you without the take.
Review the file and troubleshoot basic issues
Open the saved recording and watch enough of it to verify the picture, sound and timing. Check the start, a representative middle section and the ending. Listen for the microphone and computer audio you intended to capture, and confirm that they are not duplicated or missing. If the file is meant for editing, check that the expected tracks are available in your editing application.
If no file appears, first check the destination and whether there is available disk space. Streamlabs' troubleshooting guide also recommends checking whether your account can write to the selected directory. These are sensible first checks because a recording needs both room to save and permission to save there. Streamlabs lists additional Windows suggestions for certain errors, including updating graphics drivers, installing Microsoft Visual C++ Redistributable 2017, or running as administrator. They are not universal fixes; consult the current Streamlabs troubleshooting guidance and match a step to the error you see.
If the saved picture is blank or does not show the intended application, return to the source setup. Check that the correct display or window is selected, that the target was open and unminimised, and that the application permits capture. Try a short test with the same application after any change. A visible source label is not a substitute for checking playback.
If the image is present but the audio is missing, inspect the selected inputs and their activity. Confirm that you added the intended microphone or computer-audio source, and that you did not select a duplicate that creates echo. On Mac, revisit the system-audio routing caveat rather than expecting direct capture of other applications. If you are using separate tracks, verify the output mode, selected format and track assignments against the multitrack guide.
If playback is choppy or the computer struggles during capture, adjust quality or output resolution and test again. The vendor's suggested settings are starting points, not a promise about your processor, graphics hardware or content. Change one setting at a time and review another saved clip so you know whether the change helped.
For channels that use a prerecorded video as a continuous live source, a local Streamlabs recording is not the same as a playlist-based broadcast. This OBS playlist workflow explains that different use case. Choose the recording workflow when you need a file on your computer; choose a broadcast workflow when the channel must remain live.
Before committing, compare the operating options on the pricing page. When the file and channel are ready, start free — 24-hour trial, no card.
FAQ
Can Streamlabs Desktop record without going live?
Yes. Streamlabs documents recording as a separate option, with a Record control in the editor; you do not have to start a YouTube broadcast to make a local file. Set the recording options in Settings > Output, prepare the sources and check the saved result.
Where are my recordings saved?
The destination is the one configured under Settings > Output, so it can differ between computers and setups. Check that setting and use the completion notification after stopping to inspect or open the saved file.
Should I use a particular quality or frame-rate preset?
There is no universal ideal setting in the documented workflow. Streamlabs offers examples for different content and encoding options, but your computer's performance and the material you are recording matter; test a short clip and adjust if necessary.
Can I record my microphone and computer sound separately?
Streamlabs documents a multitrack workflow using Advanced output mode, track selection and input assignment. Its guide says up to six separate tracks are available and specifies MP4 or MKV for multitrack recording; verify the chosen tracks in the saved file before relying on them.