ChatGPT Can Now Log In to Websites Without Seeing Your Password

Back
Kevin Riedl

13 min read · 28 Aug 2026
Last reviewed

Next
Made on your device, with no Instagram connection. We copy the post link for Instagram’s Link sticker.

ChatGPT Work can now continue tasks inside supported websites after you sign in, without exposing the username or password to the model. This is not passwordless authentication. You still enter your credentials and any two-factor code, but you enter them in a secure sign-in form that sends them directly to ChatGPT’s remote browser.

The distinction matters. Hiding a password from the model removes one obvious leak path. It does not make the resulting authenticated browser session harmless. Once signed in, ChatGPT can read pages, fill forms and act with the permissions of that account. For a business, the real decision is therefore not simply “does the AI see the password?” It is “what can this session reach, change and disclose?”

Want to test a signed-in browser workflow without handing it an admin account?

 Map a Secure Pilot

What changed in ChatGPT’s cloud browser?

Availability checked 11 November 2026. OpenAI’s current browser guide documents website sign-in for ChatGPT Work on web and mobile with Plus and Pro, subject to rollout. It explicitly excludes Enterprise and Edu website sign-in. General cloud-browsing access on other paid plans or an administrator enabling browser capability does not establish access to this authenticated flow. Check the actual account before designing a company workflow around it.

The cloud browser is a separate computer in OpenAI’s cloud. It does not inherit the tabs, saved passwords, extensions, browsing history or existing sessions from your laptop. It can keep working after you close your computer, then pause when it needs a sign-in, more information or confirmation.

You do not manually select a browser for every task. ChatGPT can choose a connected app, plugin or cloud browser based on the request. OpenAI recommends using a connected app or plugin when it supports the job. That is sensible: a structured integration is usually more predictable than clicking through an interface designed for people.

How the secure sign-in flow works

  1. ChatGPT reaches a supported login page and pauses. It does not ask you to paste credentials into the conversation.
  2. A separate review model checks the destination. OpenAI says it inspects the website address and sign-in form for signs of phishing or deception.
  3. You review the domain and form. You can inspect the live website before continuing.
  4. You enter the username, password and any 2FA code. Those values go directly to the remote browser. OpenAI says the secure form keeps them outside the model and conversation. Optional saved-logins are a separate feature, described below.
  5. ChatGPT resumes inside the authenticated session. The website’s cookies can keep that session signed in for later tasks until they expire or you clear the site’s browser data.
  6. Consequential actions still require review. ChatGPT is designed to ask before actions that create a financial, legal, account or other real-world commitment.
Asset or capabilityWhat the model seesWhat still needs protection
Username and passwordNot visible when entered through the secure formPhishing review, correct domain, password hygiene and MFA
Two-factor codeNot entered into the chatThe user must verify the request and destination
Authenticated sessionCan operate pages available to the signed-in accountAccount scope, cookie lifetime, site permissions and logout
Page content after loginCan be read through screenshots and page interactionSensitive data, prompt injection and data controls
Your normal browserNot inherited by the cloud browserA separate cloud session remains until expiry or clearing

Why “the model cannot see the password” is not the same as “safe”

A password is only one route into an account. A valid session cookie can carry much of the same practical authority. If the account can view invoices, export customer data or change settings, the signed-in cloud browser may be able to reach those surfaces even though the model never received the password.

OpenAI’s Work cloud security guide treats browser access, connected apps and shell networking as separate capabilities. Website permission does not grant every account action, and content retrieved from a site remains untrusted input. Limit the account’s actual privileges and test the action boundary rather than relying on an instruction to ignore malicious page text.

This is not a theoretical edge case. NIST’s 2026 agent-hijacking analysis covered more than 250,000 attack attempts against 13 frontier models. At least one successful hijacking attack was found against every target. That study does not evaluate this specific sign-in flow, but it supports a conservative operating assumption: eventually, a browser agent will encounter content designed to manipulate it.

The right security boundary is therefore the account and workflow, not the secrecy of the password alone. Use the same design principle we apply in our AI agent security checklist: break the combination of untrusted input, valuable data and an outbound action path.

Session, saved login and task history are different

OpenAI’s dots sign-in documentation uses the same private form but separately offers optional Save to Passwords where supported. A saved login is different from a valid browser cookie; reusing it for a new sign-in requires confirmation. This does not mean the secure form puts a password in chat, nor that every account has a password manager.

To end a cloud-browser session, use Settings → Cloud browser → Browser data → Clear all. This signs you out of sites in the hosted browser. Review any saved login separately. Task conversations, hosted snapshots, Library files and memories have separate retention and deletion rules. Clearing cookies is not a purge of task records; deleting a conversation does not delete a file saved to Library. Include each retained artifact in the pilot’s cleanup check.

Cloud browser, app or custom integration: which should a business choose?

OptionBest fitMain trade-off
Connected app or pluginA supported system with a defined data or action contractLimited to the permissions and operations the integration exposes
ChatGPT cloud browserIrregular, cross-site work where no useful API existsWebsite compatibility, UI drift, prompt injection and review time
Custom API, MCP or workflow integrationRepeated, high-volume or regulated processes with measurable service levelsEngineering cost, but clearer authorization, idempotency and audit

The cloud browser is compelling for a low-frequency task that crosses several portals, such as collecting quotes or reconciling a handful of invoices. It is a poor default for a high-volume posting, payment or account-administration process. Stable production work benefits from typed inputs, narrow scopes, deterministic validation and retries that cannot duplicate an action.

Our guide to making a product usable by AI agents covers the architecture behind that custom path. The browser route is useful precisely where the structured route does not exist yet. It should not become an excuse to avoid building one for a proven, repeatable workflow.

A 10-control checklist before your first signed-in task

  1. Use a dedicated account. Do not begin with an owner, administrator, finance or production-support identity.
  2. Apply least privilege at the source. Limit the account to the records and actions the pilot actually needs.
  3. Keep website access on “Always ask.” Move to broader approval only after the domain list and task are stable.
  4. Require MFA. Treat every unexpected authentication request as a reason to stop and inspect the domain.
  5. Name forbidden actions in the task. State that the agent may prepare but not submit payments, messages, deletions or contractual changes.
  6. Keep human confirmation for consequences. Do not turn an approval prompt into a routine click. Review the target, values and final state.
  7. Separate browsing from sensitive apps. Enable only the connected apps needed for this job, and avoid mixing private inboxes with open-web research.
  8. Review data controls and retention. Decide whether screenshots and task history fit your policy before exposing customer or employee data.
  9. Clear browser data after the pilot. A password that the model never saw is irrelevant if an unnecessary authenticated cookie remains active.
  10. Measure accepted outcomes. Track successful completion, human review minutes, retries, blocked sites and incidents, not just tasks started.

What website owners should change

Treat agent identity and user authority separately. RFC 9421 HTTP Message Signatures can authenticate selected HTTP message components when the verifier trusts the key and checks the signature correctly. A header named Signature-Agent is not proof by itself. Use the current provider and edge-vendor verification instructions; signature validation does not replace TLS, account authorization or action approval.

That means “allow the signed agent” and “trust the signed-in user to do everything” must remain separate decisions. Verify the signature at the edge, preserve normal account permissions, keep rate limits, and test login, OAuth redirects, captcha fallbacks and confirmation screens. Never allow a user agent string alone or disable the WAF for a broad IP range.

Also make the workflow machine-legible. Use labels tied to fields, stable button names, explicit validation errors and a clear review step before a write. If your team is still at the discovery layer, our agent-readable website guide explains the publishing side. Sign-in support adds an action channel; it does not fix unclear pages or weak authorization.

A commercial pilot with a stop condition

Choose one reversible workflow with a meaningful manual baseline. A useful first pilot might read three supplier portals and prepare a comparison for review. A poor first pilot changes payroll, publishes customer communications or purchases inventory.

Run 30 representative cases and record task success, accepted result, intervention time, blocked-site rate, retries and any unexpected data access. Stop if the agent reaches outside the approved domains, asks for credentials in chat, attempts a forbidden action or produces an unauditable result. Scale only when the accepted cost and risk beat the manual or integration alternative.

Frequently Asked Questions

Can ChatGPT now log in to websites?
The current guide documents website sign-in for Plus and Pro on web and mobile, subject to rollout and site compatibility. It explicitly excludes Enterprise and Edu website sign-in. General paid-plan cloud browsing does not guarantee this login flow.
Can ChatGPT see my username or password?
OpenAI says credentials entered through the secure form go directly to the remote browser outside the model and chat. Optional Save to Passwords, where offered in dots, is separate from the form and from session cookies. Never paste credentials or verification codes into chat.
Is this passwordless authentication?
No. The website still authenticates you with its normal username, password, SSO and two-factor flow. The new boundary keeps the sensitive entry outside the model while creating an authenticated remote-browser session.
Does ChatGPT use the accounts already open on my computer?
No. The cloud browser is separate from your local browser and does not inherit its tabs, cookies, saved passwords, extensions or signed-in sessions. You sign in separately when prompted.
Will it work on every website?
No. Websites may block automated traffic, a login or transaction may be unsupported, and ChatGPT may restrict an action. You may need to take over or complete the final step yourself.
Should a company use cloud browser instead of an API?
Use the cloud browser for irregular or cross-site tasks where no useful integration exists. Prefer a connected app, plugin or custom API for repeated, high-volume or regulated work where scopes, validation, retries and audit need to be deterministic.
What is the biggest remaining risk?
The authenticated session. Even if the model never sees the password, it may operate pages and data available to the account. Limit the account, domain set, connected apps, session lifetime and actions before starting.

For ongoing software work rather than a single signed-in browser task, our ChatGPT Dots and GitHub bug-triage guide proposes a bounded route from report to reviewed patch, with explicit test evidence and separate shutdown checks.

Final thoughts

ChatGPT’s secure sign-in flow removes a dangerous habit: putting credentials where the model can read them. That is a meaningful improvement, not a complete security model.

Treat the cloud browser as a temporary delegated session. Give it a dedicated low-privilege account, a short domain list, human approval for consequences and a measurable task. If the workflow proves stable and valuable, move the repeated path to a structured integration with explicit authorization and audit.

Build the product, not just the backlog

If this article maps to a real product decision, Wavect can help you scope, build, harden, or lead the software work with senior founder-level judgment.

Useful service paths:

Inbox, without the noise

Follow the work that matters to you

Get a short email when we publish something new. Follow the whole blog or only the problems you care about.

What would you like to receive?
Choose your topics

Free, double opt-in, no tracking pixels.

Back
Kevin Riedl

13 min read · 28 Aug 2026
Last reviewed

Next

Get the next AI and agents field note

One concise email when we publish. No tracking pixels, and no inbox filler.

Free, double opt-in, no tracking pixels.