To upload a video to ChatGPT, open a chat, select + > Add photos & files, choose your saved video through a supported file picker, add your question, and send it. Direct video attachments are supported on eligible accounts, including Free, but availability varies by platform, upload method, and account. If Photos will not let you select the video, try Files instead.
The short version
- Saved videos: Start with + > Add photos & files in a chat.
- Picker trouble: Try Files when the video cannot be selected in Photos.
- Other upload options: Supply still frames, an existing transcript, or a separate audio file, depending on what you need analyzed.
- Live sharing: Eligible subscribers can share their camera or phone screen through Advanced Voice on iOS and Android.
| Method or option | What it does | Main condition |
|---|---|---|
| Direct video attachment | Attaches a saved video | Supported account and upload route, including eligible Free accounts |
| Files picker alternative | Attaches a video that Photos will not select | Supported Files upload route |
| Still frame upload | Supplies selected images | Free or paid account on web or mobile |
| Separate audio upload | Supplies sound without visuals | Paid subscription or workspace and an audio-only file |
| Transcript upload | Supplies the video's text | Supported document format and available upload quota |
| Live camera sharing | Shares your phone's camera view | Eligible subscriber using Advanced Voice on iOS or Android |
| Phone screen sharing | Shows what is on your phone screen | Eligible subscriber using Advanced Voice on iOS or Android |
| YouTube browser side chat | Uses a timestamped caption transcript | Supported browser, account access, and available captions |
| Companion browser setup | Configures the extension and website access | Supported desktop account and browser, subject to workspace restrictions |
| Official extension listing | Provides the ChatGPT browser extension | Chrome or a supported Chromium-based browser |
| ChatGPT Record | Captures and transcribes audio | Plus, Pro, Business, Enterprise, or Edu on macOS |
| Upload limits and storage | Helps explain upload failures | File upload access, with storage checks on web |
| Android permissions | Controls camera, microphone, and listed media access | Android labels vary by device |
| Parental controls | Controls Voice access for a teen | Linked parent and teen accounts |
| Workspace controls | Controls Voice and Record access | Managed workspace and owner access |
| Voice training controls | Controls use of shared recordings for training | Personal account with these controls |
| Official app download | Provides the supported ChatGPT app | Feature availability varies by platform and account |
Attach your saved video
Open the ChatGPT conversation where you want to ask about the clip. Select +, then Add photos & files, and choose the video through an available file picker. Add your question and send the message.
The picker route depends on your device and account. When Photos prevents you from selecting the clip, try attaching it through Files. This alternative also applies to eligible Free accounts with a supported Files upload route.
Direct attachments and live camera sharing have different eligibility requirements. You can have a supported video upload route on Free, while live camera and screen sharing require an eligible subscription on iOS or Android.
Upload screenshots or still frames
You can take screenshots on your device or use exported frames to ask about the video's visuals. These images supply selected frames rather than the complete video.
With your images ready, select + > Add photos & files and attach them. You can also drag an image into the message area or paste a copied image. Add a question about the supplied frames and send the message.
Image inputs support PNG, JPEG, and non-animated GIF on Free and paid accounts, on web and mobile. Your question should focus on the supplied images, since they do not include the video's audio or complete sequence.
Send audio or an existing transcript
For questions about the sound, attach a prepared audio-only export to a chat and request transcription or analysis. Audio uploads are available to paid subscriptions and workspaces, including Enterprise, but not Free. This route supplies the audio without the video's visuals.
Supported audio examples include WAV, MP3, M4A, FLAC, AAC, and audio-only MP4 or WebM, with audio files up to 512 MB. MP4 and WebM files identified as video are not accepted through audio uploads, so this route requires an audio-only file.
An existing transcript is another option for Free and paid accounts on web and supported mobile apps, subject to upload limits. Attach it in a supported document format, such as TXT or DOCX, and ask your question. A transcript supplies text rather than video or audio.
Share live video from your phone camera
Eligible subscribers can share live camera video in the ChatGPT app for iOS or Android. If you need the app, get it from the official ChatGPT download page. Installing a desktop app does not establish access to these mobile video features.
Open Settings > Voice and select Advanced. Start a conversation using the Voice icon, then select the camera button to share your camera view. Select the camera button again to stop.
Keep Advanced selected for camera sharing. Live does not support video or screen sharing, while Advanced remains supported. Your camera supplies a live view rather than a saved video attachment.
Show a video through phone screen sharing
Eligible subscribers using ChatGPT on iOS or Android can also share their phone screen. Showing a playing video this way is a workaround rather than a file upload. Playback-audio capture is not guaranteed, so supply a separate audio upload on a paid account or an existing transcript when you need to include the spoken content.
Select Settings > Voice > Advanced, then start Voice using the Voice icon. Open •••, select Share Screen, and complete your device's screen-sharing prompt. You can then show the playing video on your phone screen.
To stop sharing, return to ChatGPT and use the screen-sharing control. As with camera sharing, you need Advanced; switching to Live will not provide screen sharing.
Ask about a YouTube video in your browser
The official ChatGPT browser extension lets you ask about a YouTube video through browser side chat. When captions are available, ChatGPT can use the video's timestamped transcript. This supplies transcript access rather than the video file.
On a supported desktop account, open the ChatGPT desktop app and select Settings > Computer Use. Expand More browsers if needed, select your browser, then select Install to open its extension listing. You can also get the extension from the official ChatGPT extension listing.
After installation, return to Computer Use and check for Manage. Select Manage to review website access. If prompted for one-time website access, select Allow once.
Open the YouTube video, then select ChatGPT from the browser toolbar or the Extensions menu. Alternatively, right-click the page and select Ask ChatGPT, then ask about the video. On macOS, Cmd + Shift + . also opens side chat.
Side chat supports Chrome, Edge, Brave, and Vivaldi; Opera does not support it. Use one of those supported browsers, subject to your account's availability and workspace restrictions. When the video has no captions, supply an existing transcript or use a supported direct video attachment route instead.
Capture audio with ChatGPT Record on Mac
ChatGPT Record captures and transcribes audio in the macOS app for Plus, Pro, Business, Enterprise, and Edu. Using it for a video's playing audio is a possible workaround, rather than a dedicated video upload method. Record does not upload video frames.
In the macOS app, select Record. Select Stop, then Send to submit the transcript and generate notes. If required, enable ChatGPT under System Settings > Privacy & Security > Microphone and under System Settings > Privacy & Security > Screen & System Audio Recording.
Resolve upload and sharing problems
If you cannot select a video in Photos, try the Files alternative first. If an upload error gives a retry time, wait until that time before trying again. Failed uploads can count toward rate limits.
On the web, open Settings > Storage to check saved-file storage. Storage and rolling upload quotas are separate, and the remaining rolling quota is not displayed. Checking storage therefore does not show how much rolling upload allowance remains.
On Android, check permissions when camera or microphone access is the obstacle. Open Settings > Apps > ChatGPT > Permissions. For Camera or Microphone, select Allow only while using the app when offered, and review Photos and videos if listed.
Also check Settings > Security & privacy > Privacy > Privacy controls for Camera access and Microphone access. Android labels vary by device, and Google says some Pixel instructions require Android 11 or later. Permissions enable hardware access, but they do not add ChatGPT features outside your account's eligibility.
Frequently Asked Questions
How can you control whether shared Voice video is used for training?
On personal accounts with these controls, open Settings > Data Controls and turn off Include your video recordings. To disable broader conversation training, turn off Improve the model for everyone. These settings govern data use and do not enable video uploads.
Can parental controls affect Voice access?
Yes, Voice access can be controlled for linked teen accounts. In the parent's account, open Settings > Parental controls, select the teen under Family members, and review Voice Mode. Enabling Voice does not override subscription or feature eligibility.
What should you check if Voice or Record is restricted in a workspace?
Ask your workspace owner to review Voice in workspace settings. For Record, owners can open Settings > Workspace Controls > Record. Record is disabled by default in Enterprise and Edu until an owner enables it. Neither control guarantees video-attachment availability.
Should you follow an older tutorial that uses ChatGPT Atlas?
Use supported ChatGPT surfaces instead of Atlas-specific setup instructions. OpenAI has retired Atlas, and it no longer receives updates.