Guide9 steps

ChatGPT Locked Chats: how to lock a chat with a PIN or Face ID, and what it doesn't protect

ChatGPT locked chats, launched October 20, 2026, hide chosen conversations behind a 4-digit PIN or Face ID. How to lock one, which devices and plans get it, what happens if you forget the PIN, and why locking isn't encryption.

A glass PIN pad with four lit dots in front of a chat bubble whose text is faded, on a violet gradient, titled 4-digit PIN

In 20 seconds

  1. OpenAI launched locked chats in ChatGPT on October 20, 2026. A locked chat leaves your regular history, pinned chats and search, and opens only with a 4-digit account PIN or, on supported phones, Face ID, Touch ID or the device unlock.
  2. OpenAI's help center lists Free, Go, Plus, Pro, Business, Enterprise and Edu plans, on the web, iOS, Android and the new ChatGPT desktop app for macOS and Windows. Its October 20 release note names only Free, Go, Plus and Pro.
  3. Locking doesn't add end-to-end encryption, change retention or the 'Improve the model for everyone' setting. Resetting a forgotten PIN permanently deletes every locked chat, and OpenAI Support can't recover them.
Steps

ChatGPT locked chats let you hide chosen conversations behind a 4-digit PIN or, on a supported phone, Face ID, Touch ID or your device unlock. OpenAI announced the feature on October 20, 2026, and is rolling it out gradually across ChatGPT. It's built for one situation: someone else looking at your screen or holding your phone. It isn't a vault. A locked chat isn't end-to-end encrypted, OpenAI keeps it as long as any other chat, and if you reset a forgotten PIN, every locked chat is deleted for good.

This guide is based on OpenAI's help article "Using locked chats in ChatGPT," its ChatGPT release notes, its Lockdown Mode help article and TechCrunch's launch report, checked October 23, 2026. Locking also leaves ads alone: OpenAI says ads can still appear in locked chats where they're available, which our explainer on ads in ChatGPT covers.

The short version

  • What: a "Lock chat" option that moves a conversation into Projects > Locked chats, out of regular history, pinned chats and search.
  • How you open it: a 4-digit account PIN, or Face ID, Touch ID or Android device unlock on supported phones.
  • Where: the web, the iOS and Android apps, and the new ChatGPT desktop app for macOS and Windows.
  • Forget the PIN: resetting it permanently deletes all locked chats and their protected files and memories.
  • What it doesn't do: encrypt the chat end to end, shorten retention, opt you out of model training, or remove what ChatGPT already remembered.

How do I lock a chat in ChatGPT?

Open the chat, tap or click the more options menu (•••), select Lock chat, and confirm. The first time, ChatGPT asks you to create a 4-digit PIN and enter it twice. That's the whole process, per OpenAI's help article.

To start a new conversation that's locked from the beginning, go to Projects > Locked chats, enter your PIN or use device authentication, and type a message. OpenAI suggests locking an existing chat first if you've never used the feature.

To open a locked chat later, go to Projects > Locked chats, authenticate, and pick the chat. That gives temporary access; it doesn't unlock the chat permanently. Access ends on its own:

  • Web and desktop: when you refresh, close the tab, or leave the tab or window for a while. A new tab needs the PIN again.
  • iOS and Android: after 1 minute away from the locked chat, or 1 minute with the app in the background.
  • Anywhere: select Lock now to end access immediately and hide the titles in the sidebar.

To unlock a chat for good, open it, select ••• and Unlock chat (just Unlock on iOS), then confirm. It returns to your regular history, not to any project it came from.

If you don't see Lock chat, OpenAI says the feature may not have reached your account yet.

Which devices and plans get locked chats?

Locked chats work on the web, in the iOS and Android apps, and in OpenAI's new ChatGPT desktop app for macOS and Windows. OpenAI's sources disagree slightly on plans, so check the table.

ItemWhat OpenAI saysWhere
PlansFree, Go, Plus, Pro, Business, Enterprise and Edu; not ChatGPT for HealthcareHelp article
Plans (launch note)"Rolling out gradually to Free, Go, Plus, and Pro users"Release notes, October 20, 2026
PlatformsWeb, iOS, Android, new desktop app for macOS and WindowsRelease notes and help article
BiometricsFace ID or Touch ID on iOS; "Use device unlock" on Android; desktop uses the PINHelp article
Voice in locked chatsWeb, iOS and Android; not the desktop appHelp article
RegionsNot statedNeither source

Table 1The help article lists seven plans, the launch note four; neither names regions. Source: OpenAI Help Center, "Using locked chats in ChatGPT," and ChatGPT release notes (October 20, 2026).

Last verified: October 23, 2026

TechCrunch's October 20 report lists the same seven plans as the help article. For what each plan costs, see our ChatGPT plans comparison.

Which chats can't be locked?

The Lock chat option doesn't appear for temporary chats, archived chats, group chats, chats created from a shared link, chats with scheduled tasks, chats with meeting recordings, chats tied to background tasks, chats whose responses are saved as project sources, and chats in the earlier Health experience. Codex conversations and local Work conversations can't be locked either. To lock a chat with a scheduled task, OpenAI says to delete the task first, including a paused one.

Locking a chat inside a project removes it from that project, so it stops using the project's instructions, files and memory.

What happens if I forget my ChatGPT PIN?

Forgetting it deletes nothing. Resetting it deletes everything locked. OpenAI's help article says a reset "permanently deletes" all locked chats, memories stored only within locked chats, and files protected only by the old PIN. It can't be undone, and OpenAI Support can't recover the content. A workspace administrator can't reset your PIN or unlock your chats either.

To reset: Settings > Security and login > Account PIN, then Reset (web or desktop) or Reset PIN (iOS or Android). You verify your identity and confirm a button that names the deletion, such as "Reset PIN & Delete" on the web or "Delete all and reset" on iOS.

Changing a PIN you still know is different: same settings page, Change, enter the current PIN, set a new one. Your locked chats stay.

"ChatGPT locked me out": after too many wrong PIN attempts, OpenAI says access is temporarily disabled until a retry period ends. Waiting keeps your chats; resetting doesn't.

The PIN belongs to your OpenAI account, not a workspace, and works on every device signed in to that account. Face ID or device unlock is set up separately on each device and never replaces the PIN.

Does locking a chat encrypt it?

No. OpenAI's help article says locking "does not create end-to-end encryption or a separate encrypted vault." The chat is stored on OpenAI's servers the same way as before; the lock controls who can open it in the app.

Locking doesLocking doesn't
Hide the title and content from regular history, pinned chats and searchAdd end-to-end encryption or a separate vault
Require a PIN or device authentication to openChange how long OpenAI retains the chat
Hide files uploaded or created in the chat from SpaceChange your "Improve the model for everyone" setting
Keep new memories inside the locked areaRemove memories saved before you locked the chat
Block sharing, archiving, pinning and movingDelete an existing shared link, or stop ads

Table 2Locking controls who can open a chat on your devices, not how OpenAI stores or uses it. Source: OpenAI Help Center, "Using locked chats in ChatGPT."

There are practical gaps too. Other open devices may keep showing a chat's title until their list refreshes, and on iOS previously loaded content can stay visible until the app refreshes, so OpenAI suggests checking other devices before you share your screen. Anyone who opens a locked chat can still copy its contents. In ChatGPT Enterprise and Edu workspaces, locked chats remain in the message logs available through OpenAI's Compliance Platform, and admins can't turn the feature off.

Training: locked chats follow your existing "Improve the model for everyone" setting. If it's on, locking doesn't turn it off. If you want a conversation kept out of history and training, OpenAI's tool for that is a temporary chat, which can't be locked.

Memory: if Memory is on, a locked chat can use your regular memories. Anything it learns after locking is remembered only for other locked chats and doesn't carry into regular chats. Memories saved before you locked the chat stay where they are. After you unlock a chat, the whole conversation may feed regular Memory if your settings allow it.

Search: locked chats drop out of regular search, though OpenAI says results can take time to update. You can't search across all your locked chats.

Shared links: locking a chat that already has a shared link doesn't delete the link, and content already shared stays available; new messages aren't added. To stop access, delete the link separately. Shared links can outlive a conversation's privacy, as Meta's Muse share links showed in October 2026.

Is ChatGPT Lockdown Mode the same thing?

No. Lockdown Mode is a separate, older setting, opened to all logged-in users on June 4, 2026, according to OpenAI's release notes. It isn't about who can see your chats. It limits ChatGPT's access to the web and outside services to reduce the risk of data being stolen through prompt injection.

With Lockdown Mode on, OpenAI's help article says browsing is limited to cached content, deep research and agent mode are disabled, ChatGPT can't download files for data analysis, and live connector access is restricted on personal accounts. You turn it on under Settings > Security > Advanced security. OpenAI says it is "not intended for everyone," doesn't stop injected text from affecting answers, and doesn't guarantee data can't leak. It also leaves memory, sharing and training settings unchanged.

What's confirmed, reported and uncertain

  • Confirmed (OpenAI): the October 20, 2026 launch; the steps; PIN and biometric options; platforms; the list of chats that can't be locked; that a PIN reset deletes locked chats permanently; and that locking doesn't add encryption or change retention, training or existing memories.
  • Conflicting (OpenAI's own pages): the help article lists Free, Go, Plus, Pro, Business, Enterprise and Edu; the release note names Free, Go, Plus and Pro. TechCrunch's report matches the help article.
  • Uncertain: regional availability, which neither OpenAI page states, and how long the gradual rollout will take.

Read next

Comments (0)

More on ChatGPT

View all
Products · CompareChatGPT vs Perplexity (October 2026): plans, prices, search, models and which to pickChatGPT vs Perplexity in October 2026: Perplexity Pro lists $17 a month billed annually and Max $200 a month, against ChatGPT Plus at $20 and Pro from $100. How search and citations, deep research, agents, models, file limits, ads and privacy compare, from both companies' own pages.·8 min read
Products · CompareChatGPT vs Microsoft Copilot (October 2026): plans, prices, models, Microsoft 365 integration and which to pickChatGPT vs Copilot in October 2026: ChatGPT Plus costs $20 a month; Microsoft bundles Copilot into Microsoft 365 from $9.99, with Premium at $19.99 and a new $99.99 Pro plan. Models, Office integration, business pricing and data protection, from both companies' own pages.·8 min read
Models · CompareGPT-6 Astra vs GPT-6.1 Sol vs GPT-6 Luna: what each OpenAI model is for, prices, limits and which to use (October 2026)GPT-6 Astra vs 6.1 Sol vs Luna, compared on OpenAI's own pages as of October 23, 2026: Astra costs $10/$50 per million tokens, GPT-6.1 Sol $2/$10 and Luna $0.10/$0.50, all with a 1.05M-token context, but they sit in different parts of ChatGPT.·8 min read
Products · Analysis"Publish fake stats, get cited in LLMs": how AI answers get gamed, and what AI search companies sayA Peec AI researcher says 50,000 unchecked, AI-generated statistics pages drew 10,000+ ChatGPT sessions and links from 330+ sites. What the 'fake stats cited in LLMs' experiment shows and doesn't, what the research on gaming AI answers found, and what OpenAI, Google and Perplexity say about sources.·9 min read
The Week in AI · every Friday

New guides and explainers, every Friday.

The week’s biggest stories in AI and tech, what they mean, and what to read next.

Free. One email a week. Unsubscribe in one click.