@user-f43a29 , @user-b43477 Sorry to disturb but I am interested in something similar. Would it be possible to share what solution/ method was developed?

@user-f43a29 Thank you. Just a rough ambient measurement would suffice, thank you..

Thank you. If we just want to determine regions where luminance changes (to evaluate pupil diameter), would changing the scene exposure post-hoc (as you suggested) help?

@user-f43a29 Thank you. Yes, it is indoors and I shall investigate that. I think integrating the sensor might also be useful in sitatuons where there is a sudden change in light (a time lag in motion activated light)

hi! I had a quick question; I've been using Neon with typical participants (and I've been loving it!), and next week I'm talking with an Ophthalmological research group who doesn't have eye tracking experience; does Neon have known problems recording data on participants with Cortical Visual Impairments (CVI's), or atypical eye alignment/movements more generally? I know algorithmically, you're using ML to get pupil position --> gaze, but I wasn't sure how damaged this estimation process would be by atypical eye movements.

Hey, we are using the Neon glasses and today the Pupil Companion on the phone crashed during a recording. In the list of recordings on the phone it shows all the (successful) and uploaded recordings; some before, some after the crash which all show up in the browser. In the bottom it shows one recording as "Queued" having the (incorrect) timestamp of the preceding successful recording. After logging in and out of the app we got a notification that there was an unsaved recording (with the correct time stamp). After clicking "Save" nothing noticable happened. If you want to play the recording it fails with "Video playback not possible" (translated from my local language). Any idea how to recover that video?

Thanks. I was just told that they managed to upload the video after rebooting the phone. The processing looped some time between 47% and 54% but eventually run through. Just logging out and back was not sufficient. We also experience (subjectively) more crashes where the video suddenly went dark (with videos being around 28 - 30 minutes long). Should I still open a support ticket? The original problem of uploading the aborted video is solved.

Hi, I was wondering if the Pupil Labs Neon has an ambient light sensor and whether it is possible to extract the data from it. Thank you very much.

Hi Miguel @user-d407c1 , I am trying to get in touch about the analyses you had helped me with a few weeks ago. It seems like a simple fix but I am having some difficulty and am wondering if you take a look at the issue. It would be another huge help!

Hi @user-480f4c ! Oh I am sorry for forgetting about that. I hope everyone enjoys the break. Thank you!

Hi Rob, no it was kept active during the recordings

Does pupil accept returns for Neon if I try the module out and it doesn't end up integrating well in my setup?

Hi @user-e3fdf5 ! I have replied to you via email.

Hi - is there a best recommended way to cut a recording in a way that the videos and sensor data are cut in the same place? For example, I have a video that was stopped twenty minutes later than intended, so I want to remove the last twenty minutes from the recording across all sensors. Thanks!

I work with the python library so pl-neon-recording typically. In my case, I would want to discard the original recording based on certain timestamps (because it could include personal identifiable information). But assigning events sounds like a promising approach

Ok I will check that out thank you!

I just met with a neuro ophthalmologist at Cincinnati Childrens Hospital, he'd like to talk to a product specialist who can help him order a neon -- and maybe other things? Is there an email I can give him?

so is this project compatible with vive headsets through 3d printing?

Hi @user-0f152a 👋 ! If by “project” you mean product, then yes, Neon could be used in XR/VR setups. We offer turnkey integrations for headsets like the Quest3 and Pico4. But you could also get a Bare Metal and build your own mount to integrate with virtually any headset.

Is this 7 thousand dollars?

Im just looking to slam it into a vr headset and play vrchat, but i already gave up and bought some parts that ill probably just end up breaking 🙃

Hi, I just have a quick question - what is the best way to switch a datasets workspace on Pupil Cloud? We have a dataset that was initially recorded in the wrong workspace.

Also is there a way to reupload a downloaded dataset to Pupil Cloud after it’s been deleted?

Thanks for your help!

Thanks @user-f43a29, it’s on the same account, but would it theoretically be possible to transfer to a new account and then transfer back into the intended workspace?

Ok, I’m just trying to figure out what I can do with the datasets that were erroneously recorded in the wrong workspace (but we still want to utilize group level pupil lab analysis tools).

Mainly just exploring the mapper tools - interested in AOI / dynamic AOI analysis

Hi @user-1beb67 , what version of Neon Player are you using? It has been updated since then, so you might need to download the new version first.

Thanks. Would you be open to sharing the recording with [email removed] so that we can take a closer look?

Hi I have a question about the recordings. For our research we make recordings that include footage that is sensitive. Therefore I was wondering if the recordings are uploaded to the pupil cloud, can I delete them from the phone? Or will this delete the recordings from the cloud? And if I delete the recordings from the cloud, will this automatically delete the recordings from the phone?

Hi! I’m trying to export blink data from some Neon recordings that were made before Companion v2.9.0-prod. From the docs (https://docs.pupil-labs.com/neon/neon-player/getting-started/), it sounds like older recordings should be opened with Neon Player v4.x, but I wasn’t able to find a public v4 download on the GitHub releases page (https://github.com/pupil-labs/neon-player/releases).

I wanted to ask:

1. Is there an archived/downloadable Neon Player v4.x build for Linux that you’d recommend for these older recordings?
2. Is there any way to export blink data in batch for multiple recordings at once, similar to the ”Project“ function on Pupil Cloud?
3. If not in Neon Player, is there another recommended workflow for batch blink export for older recordings?

Thanks a lot!

Hi !! We are plan to buy Pupil Labs Neon eye-tracking glasses the "I Can See Clearly Now" variant. My question is: what is the procedure for applying for an academic discount? Do we need to provide proof of university affiliation?

Hi, I was wondering if there's a specific reason the top part of the neon module is exposed in some frames but covered in others?

Yup, I wondering if the nothing to see frame blocks any features that the others have. I just noticed all the other frames have that top part exposed, so I was wondering why.

Alright, What are the specific differences between the academic version and the standard version of Neon? I would like to understand whether there are any differences in hardware, software access, data export capabilities, or included accessories, or if the distinction is based solely on pricing. Is full payment required before shipment, or do you offer alternative options such as cash on delivery.

Thank you so much for your quick reply! That's clear. We recently purchased the unlimited analysis plan for our device for 3 years. However, we did not purchase the face blurring and audio removal. Is it still possible to add this?

@user-f43a29 Thank you for your prompt reply !!

Thanks! One more question, is the face blurring an option you can turn on and off? Or does it do this automatically for all recordings when this is part of your Unlimited Analysis Plan?

hi! I am using the neon eyetracker to collect data. the pupil cloud fills up pretty quickly, and for some participants I forgot to download old files from pupil cloud and so I did not have enough storage for the new participant data. for these participants, therefore, I manually had to transfer recording data from the neon companion device to my computer. the data I downloaded from pupil cloud was the 'Timeseries data' only. the data I manually transferred was the Native recording data. now here comes my question: these data files seem different (different filenames and extensions). the main data file of importance for me is the one called '3d_eye_states.csv' from the Timeseries only data. how do i get this data from the Native recording data? thank you very much in advance for the help!

Hi @user-525b37 , you can use this Python example script for our pl-neon-recording library to export the CSV files from the Native Recording Data.

Dear team, none of the videos is running in Pupil Cloud today. Is there a technical problem in your servers?

Hi @user-3c26e4 , recordings are playing back here on my end. Could you open a Support Ticket in 🛟 troubleshooting ?

Hi, Is it possible to stream real-time gaze data from Neon into Unity (e.g., via SDK, API, WebSocket, or UDP) with low latency (Does Neon support real-time gaze streaming (not just offline recording))?

Thanks, that’s very helpful. To clarify, we are planning an XR-based experiment in a controlled VR teleoperation environment. Our goal is to use Neon’s real-time gaze stream in Unity to estimate whether the user has visually acknowledged specific AOI regions real-time!

Hi all, another question. My neon companion device displays a low storage warning, stating that there's < 30 min free space to record. I've deleted almost all recordings from the neon companion app folder, but the warning is still there. What could be causing this problem? Thanks a lot!

I have the Just Act Natural frames, is there a solution for interchangeable prescription lenses for these frames? I would have multiple wearers with different prescriptions and would need to swap between lenses frequently. Assuming I had an optician make lenses (is this the best way?), is there a fast way to swap, or snap-on frames that would fit?

Im trying to stream over LSL from the companion app but Lab Recorder can not detect any stream.

Both are on the same network and I have turn stream over LSL on in the app.

Any advice?

Im using my home WiFi

Has anyone managed to stream the eye using the pupil neon without issues? If so, how did you do it? I think this feature should be added to neon.local:8080

Hi! I have 2 questions regarding the PupilLabs Neon glasses and the Cloud.

First, I would like to ask whether it is possible to connect the glasses to a phone through an adapter (e.g., a USB-C splitter). Our goal is to keep the phone connected to a power source, such as a power bank, during longer recording sessions. We are unsure whether using such an adapter might interfere with data transfer or recording performance, and would appreciate your guidance on this.

Secondly, we are currently using the basic cloud storage capacity (approximately 3 hours of footage, if I'm not mistaken). Could you please clarify what happens once this limit is reached? Are we still able to continue recording, and if so, where is the data stored? Or does the system stop new recordings when the storage is full?

Thank you in advance for your help!

I want to know how long it will take to process a video of about 10 seconds using enrichments, because I've already been waiting for 4 hours, and it feels like there's no end in sight.

@user-f43a29 I really want to know — the enrichments have been running for a very long time. Are there any solutions

@user-f43a29 I need your help

Hi @user-6ccf3a ! The time it takes to complete depends on how big the temporal selection is and how many other users are running Enrichments on Pupil Cloud at the same time.

If it is not finished by later today, please write again and I will ask the Pupil Cloud team to check.

Hi @user-6ccf3a! Is it just one video of 10 seconds you're trying to process? Could you please share some more details, e.g. which enrichment you're using, is it just on a single video etc? Additionally, have you tried refreshing the browser tab as sometimes it can cache the page and appear like the enrichment hasn't finished, when in fact it has.

@user-4c21e5 Thank you . Is this image clear enough? There's some Chinese text, but it shouldn't affect readability. I captured about 10 seconds of the image, and it's been processed for 5 hours. I'm not sure if it's a problem with my settings.

@user-4c21e5 My previous account was banned, so I'm using this account to continue giving feedback.

@user-f43a29 Do you (or anyone else here) have code we could use to set it up so we don't have to stumble through it?

@user-f43a29 Thank you, I think for now we should be able to give it a shot. I am just glad we don't have to reinvent the wheel and probably land at a worse implementation than if we didn't ask

Thank you @user-f43a29 ! I hope you have a lovely weekend!

Hi @user-f43a29 ,

I recently ordered the Pupil Labs Neon eye-tracking glasses (“I Can See Clearly Now” variant), and I wanted to ask how long the processing and delivery usually take.

I have a flight scheduled for May 4th, so I would like to know if the order is likely to arrive in Belfast, UK before that date.

I have already contacted sales@pupil-labs.com but haven’t received a reply yet.

Could you please provide an update on my order status?

Thank you.

@user-f43a29 On 04/17/2026

hey, is there a citation for the Companion app? I couldn't find a proper paper to cite it in this webpage https://docs.pupil-labs.com/neon/data-collection/publications-and-citation/

Hi @user-1beb67 👋 ! There isn’t a specific citation for the Companion App itself. When citing Neon, we generally recommend referencing the accuracy whitepaper and/or the accompanying publications on pupil detection or fixation detection, depending on what is most relevant to your work.

If you still need to reference the app, the typical approach is to include: - the app name - the version used - the date of use

You can find this information in the recording metadata (info.json), even if the app has been updated since.

Here’s an example in APA style:

Pupil Labs. (2024). Neon Companion App (Version 2.X.X) [Mobile application]. Used on [date]. Ref.

Hi, pupillabs team. I tried to run your script given here, but it doesn't work. Everything is up to date, but it says

Hi Pupil Labs team! I've been using the Neon realtime API to stream the RGB scene image with gaze overlay and usually it works fine! But I sometimes receive frames that seem to be degraded or corrupted somewhat, attached here. Usually when I see this starting to happen more frequently, I restart the android phone and start up the Companion App again to return to normal images. I was wondering if you might have any advice on why sometimes frames like this occur and how to prevent them, if possible? Thank you very much!

Hi @user-d407c1 Got it, thank you so much for the fast response! Yes, I use an institutional (university) Wi-Fi network. I think it should be 5GHz, but yes true, probably quite congested since it is institutional (university) Wi-Fi 😅 In the case where I may need to continue using the institutional Wi-Fi, would you happen to have any advice or tips? Maybe I can try having my computer slow down the stream ingestion a bit? Thank you!

Hi Pupil Labs team,

I’m experiencing an issue with uploading recordings to Pupil Cloud and was hoping you could help.

I created a new workspace because I want to use the “blurred faces” and “no audio” features. Recording with my Pupil Neon works fine, and the recordings upload from my phone without any issues (and as expected, they are not yet blurred at that stage).

However, once the recordings are uploaded to Pupil Cloud, I can see them uploading, but I’m unable to open them afterward. After some time, I receive the following error message:

“Video transcoding failed for this recording. We have been notified of the error and will work on a fix. Please check back later or get in touch with [email removed]

I waited for about half a day and also tried uploading a new recording, but the issue persists.

What’s notable is that when I record and upload in a different workspace (with face detection and audio enabled), everything works fine and the videos process without any problems.

Does anyone know what might be causing this issue or how I can resolve it?

Thanks in advance for your help!

Strange, because I have the newest version of the app and also its enabled

Hi pupil Labs Team! I recently used the Neon Eye Tracker and wondering if there is any way to turn off the scene camera (RGB) via software? If so, can we still record the eye movement data while the scene camera is turned off? Thank you!

Hi again, Is there any eye tracking varient that has IR scene camera rather than RGB ?

Hi @user-752d4a 👋 ! Could you share a bit about your use case? Currently, you can't disable the scene camera via software unless a VR frame is detected.

And no, we do not have variants with IR cameras at this moment.

Adding to this, you can stream the data and in that case the scene camera would be off until you start that specific stream.

Hello pupil labs team! I am using the real time api to connect the running device to the glasses. This process is done using WiFi. Is there a way of connecting using the glasses' usb-c?

Hi @user-250d7c 👋 ! While not directly as the Companion Device acts as battery, storage, and computing unit, you can use a USB-C hub to connect it via ethernet. See https://docs.pupil-labs.com/neon/hardware/using-a-usb-hub/

Hi @user-752d4a , just adding to my colleague, @user-d407c1 's, response, if you can rigidly fix an IR camera to the wearer's head, say with a helmet of some sort, then you can do a mount calibration, similar to what we do for Neon XR, to convert the gaze data to its coordinate system.

hi! I would like to ask help because I'm trying to edit the videos but when I hit visualization it is showing "error". Is there a reason why this may be? This didn't happen before. Thank you

Hi @user-c3bb4a , the Pupil Cloud team is aware and is working on resolving this. We will update you as soon as it is resolved.

I appreciate it

Thanks for sharing the context, I will forward it to the team. And yes that's correct, in streaming mode, only requested sensors are turned on.

Hi there. Tell me, please, how to upload the data into the pupil cloud if I am going to use the free plan? I have created the workspace and wanted to import the files into this workspace. The button "create a project" is inactive. When I drown the data in the workspace black field, nothing happens. My colleague gave me data on a hard drive to work with.

Hi @user-ee74ae , it's not possible to upload recordings to Pupil Cloud from an external harddrive. The upload must happen from the phone that made the recording. Also note that it is not possible to remove recordings from the phone and then later copy them back for upload. This is done for data integrity reasons.

What are you trying to analyze exactly? Note that you can use Neon Player with the data, if it is in the Native Recording format (i.e., the raw data from the phone).

Hello once again pupil labs team! I am trying to understand what's the best way of detecting when an eye is closed or open (individually). Throughout my soft research, I experiment things like the BlinkEventData (which is not very useful nor practical to use in my context) and receive_gaze_datum().eyelid_aperture_* and receive_gaze_datum().pupil_diameter_* which are better, but like mentioned in an old comment on this chat, is very dependent on the user and not consistent. So my question is if there is any other reliable way of doing this?

Hi @user-250d7c , which old comment are you referring to? The eyelid aperture functions would basically be the way to go about this with Neon. We even have an Alpha Lab guide in this vein.

Maybe I dreamt about the part of using BlinkEventData... But yeah, I was following this comment here!

Either way, thank you for your quick response

Hi @user-ee74ae , yes, that is possible. They can add you as a Collaborator on their Workspace.

Hello, can I ask is there are limitations of trial in Pupil Cloud? Is it limiting the storage or period? Also, how can we apply educational discount for it? Thank you.

Rob. One more question. I can analyze some data offline by Neon Player. But I couldn’t not see in this application the visualizations. Is it possible to see and analyze visualizations (heatmap, scan passes) offline?

Thank you. Does that mean that we have 2 hours of analysis per session or a total of two hours for the entire analysis? In other terms, is the 2 hour quote for continuous use or total analysis time? If we choose to use the Pupil Player - can we conduct enrichment analyses, such as AOI and manual reference mapper using the Pupil Player, or is that only supported on Pupil Cloud? Thank you.

@user-ee74ae , Neon Player does not offer all the features of Pupil Cloud, but you can make heatmaps with it. You need to use the Surface Tracker plugin to do that.

Neon Player currently does not produce scanpath visualizations. If you want to make them yourself locally, then you can adapt the code from this Alpha Lab guide to use the pl-neon-recording library.

Hi all, I am having issues connecting to the pupil neon glasses. I am trying to synchronize data with stimulus presentation. Using matlab the when calling Device(), 'Device found!' is returned, but then the script is stuck. I tried using python directly (from pupil_labs.realtime_api.simple import discover_one_device and device = discover_one_device()). I can connect to the device and see device.phone_ip, but not device.serial_number_glasses or device.serial_number_cam. However, on the companion device, both serial numbers are showing. Would you have any guidance for troubleshooting?

Still troubleshooting and was told by support that for the neon glasses 'device.module_serial' is the right attribute to get the serial number and that does show up. The issue with my glasses then is when I try to get gaze data with device.receive_gaze_datum(). I get this error: "Raw gaze data has unexpected length: [email removed] timestamp_unix_seconds=1777329650.3750148) Traceback (most recent call last): File "/home/hbml/.local/lib/python3.10/site-packages/pupil_labs/realtime_api/streaming/gaze.py", line 410, in receive cls = data_class_by_raw_len[len(data.raw)] KeyError: 105 Stopping run loop for rtsp://10.167.49.28:8086/?camera=gaze&audioenable=on"

Hi @user-c78419 , so , serial_number_glasses and serial_number_cam are for our previous eyetracker, Pupil Invisible. For Neon, you want to check the module_serial attr.

With respect to that error, you need to update your copy of the Python Real-time API package. You can do that as follows:

pip install --upgrade pupil-labs-realtime-api

Regarding the MATLAB issue, could you share the contents of MATLAB's console when it hangs? How long do you wait when it hangs? You may need to wait about 30 seconds the first time, since Python in the background may still be byte compiling any new packages.

Hi! We have started using PupilLabs crawl walk run with children recently. The children wearing the glasses complained that the sensor area on the nose and forehead is heating very quickly. They are able to keep in on maybe for 15-20 minutes maximum and then they reported feeling very uncomfortable. Do you have any suggestions for managing this?

Foam pads are attached to the frame....heatsink, I do not know. Could you show me what it looks like?

yes it has there..

children can be quite sensitive so we have to think of ways to handle this issue without compromising the structure of the glasses that much

Hi team, I recently started to use https://github.com/pupil-labs/pl-realtime-tui . It works great. However, the current red/green status cues are difficult for me to distinguish reliably during use because I have a common male red-green color vision deficiency. I would like to request an accessibility improvement. e.g., additionally using text labels or distinct shapes/symbols, using color-blind-safe palette by default etc.,

Hi everyone! I have the new version of the Neon Player and I noticed that there’s an option to visualize the corrected fixations for optical flow. However, once I download the file, what data do I actually get? the corrected one? In theory, based on the algorithm you use to extract fixations, what is labeled as “fixation” might also include pursuit, is that correct? Thanks a lot for the clarification

Hi! Sorry if this question has been answere before, but I was wondering whether the output of the magnetometer allows for extraction of azimuth relative to true north? Essentially my use case is that I would need to use it as a compass to relate people's field of view to the surroundings on a map, without having to manually recalibrate to north constantly. Thanks 🙂

are there any resources where there are general instructions for creating events on pupil cloud and then extracting fixation data from specific events indicated in the eye tracking recording video?

we have not had the onboarding call yet

thank you, I did that just now