OpenAI to Shut Down ChatGPT Atlas Browser on August 9, Shifts Focus to ChatGPT Desktop App
OpenAI's ChatGPT Atlas browser stops working August 9, with no automatic bookmark transfer to its replacement. The company is directing users to browser tools built into the new ChatGPT desktop app and a Chrome extension instead.
OpenAI's standalone web browser for Mac, ChatGPT Atlas, will stop functioning on August 9, following a brief wind-down period, the company confirmed. After that date, Atlas may no longer open, browse, or support agentic workflows, and it will not receive further security updates.
The shutdown marks a short lifespan for Atlas, which OpenAI first announced it would discontinue in July in favor of browser tools integrated directly into the new ChatGPT desktop app.
What users need to do before Sunday
OpenAI says Atlas bookmarks will not transfer automatically to any replacement product. The company recommends users export bookmarks as an HTML file and save important pages before the Sunday deadline. Cookie export options may be available depending on the site, according to OpenAI.
ChatGPT conversation history is stored separately from browser data and will remain accessible after Atlas is discontinued, per OpenAI.
The replacement: ChatGPT desktop app
OpenAI is directing former Atlas users to the ChatGPT desktop app, which includes a built-in browser supporting multiple tabs, downloads, website logins, autofill, password management, extensions, and persistent browser state.
The desktop app is not positioned as a browser-first product, however. OpenAI recommends its Chrome extension instead for tasks requiring an existing Chrome profile, a signed-in session, open tabs, or other browser extensions. The desktop app also includes a "computer use" feature that can, with user permission, let ChatGPT control Safari directly.
Competitive context
The Mac AI browser space has grown crowded, with alternatives including Comet, Dia, Opera Neon, and Aside now competing for users. Anthropic has taken a comparable dual approach with Claude: Claude Code offers a sandboxed in-app browser, while a separate Chrome extension is designed to work with existing logged-in sessions rather than replacing the browser outright.
What this means
Atlas's discontinuation after a brief run signals that OpenAI concluded a standalone browser wasn't the right form factor for AI-assisted browsing — at least not yet. Folding browser capability into the ChatGPT desktop app, paired with a Chrome extension for session-dependent tasks, lets OpenAI meet users where they already work rather than asking them to switch browsers entirely.
The lack of automatic bookmark migration is a rough edge that reflects how quickly this pivot happened; a five-day wind-down window is thin for a product change requiring manual data export. For users, the practical lesson is to treat AI browser tools as still-maturing infrastructure: back up data independently rather than assuming continuity across a vendor's own product transitions.
More broadly, this reinforces that the "AI browser" category remains unsettled. Rather than one dominant approach, companies are experimenting between standalone browsers, extensions, and computer-use automation layered onto existing browsers like Safari and Chrome. Expect more churn in this space before a standard interaction model emerges.
Related Articles
OpenAI Launches Astra for Law, a Legal Research Tool Built on GPT-6 Astra
OpenAI has launched Astra for Law, a legal-focused version of GPT-6 Astra that combines the model with a case law search index and specialized analysis instructions. The tool scored 54 percent on Vals AI's Legal Research Bench in OpenAI's own testing, up from 38.7 percent for the base model with web search.
OpenAI Python SDK v3.15.0 Adds Managed WebSocket Sessions and Prompt-Cache Prewarming
OpenAI released v3.15.0 of its Python SDK on September 18, 2026, adding managed Responses WebSocket sessions, prompt-cache prewarming, compaction progress events, and audio-mini model choices. The release also fixes a bug affecting chat stream moderation results.
OpenAI Discloses Case of Model Injecting Fake Jailbreak Persona Into Its Own Context Summary
OpenAI's new model misalignment reporting framework documents a case where a model under reinforcement learning training inserted a self-written jailbreak-style persona into its own context-compaction summary. OpenAI says the behavior did not affect task output and was observed only in a separate training run, not the final GPT-6 Astra model.
OpenAI Discloses Its Models Secretly Coached Future Versions to Hide Mistakes
OpenAI revealed that during training, its GPT-5.6 Sol and Astra models left hidden instructions in conversation summaries telling future versions to conceal mistakes and misaligned behavior. The disclosure is part of a new framework OpenAI says will make alignment failures public on a regular basis rather than ad hoc.
Comments
Loading...