ChatGPT Can Now Send Texts for You with New Apple Messages Plug-in
On August 20, 2026, OpenAI officially launched an Apple Messages plug-in for iPhone and Mac users. With this plug-in, users can connect their Messages inbox to ChatGPT, letting the AI help sort, analyze, and edit messages — even drafting and sending texts on your behalf. It is another major move by OpenAI into the Apple ecosystem following its deep Siri integration, and it once again puts the 'AI and personal data' privacy topic in the spotlight.
What the Plug-in Can Do: From Sorting Messages to Sending Texts
According to OpenAI, the plug-in's core value is handing your 'digital conversations' over to AI. After connecting the Messages inbox, users can sort, analyze, and edit their messages directly inside ChatGPT. The plug-in also works with Codex and ChatGPT Work, so users can rely on it not just for personal use but also in professional work scenarios.
In an official commercial, a user asks ChatGPT to suggest follow-up messages to contacts based on messages received the previous day. You can also ask ChatGPT to delete messages, draft and send messages on your behalf, or search for information buried deep in your message history. In other words, almost the entire message management workflow can be handed over to AI.
Privacy Concerns: Local Processing and the No-Index Promise
As with most AI-related features, this new capability immediately raised privacy questions. Letting AI touch your entire message history means it can read the most intimate conversations you have with family, friends, and colleagues. OpenAI told Bloomberg that the plug-in runs locally on a user's machine and that it 'doesn't create an index of all someone's messages.'
However, the exact technical details of 'local processing' and 'no index creation' are not yet fully clear. Local processing can reduce the risk of data upload to some extent, but it remains unclear when messages are sent to OpenAI's servers and which operations trigger cloud processing. TechCrunch has reached out to OpenAI for more information.
OpenAI's Safety Warning: Don't Turn On Persistent Approval
When it comes to sending messages on your behalf, OpenAI is notably cautious. The company explicitly encourages users to keep an eye on what ChatGPT is doing and discourages enabling persistent approval mode. OpenAI writes that doing so 'removes your final chance to review a message before ChatGPT sends it as you.'
This warning is crucial: imagine AI misinterpreting your intent and sending an ill-phrased message to your boss or an important client without your confirmation — the consequences could be quite embarrassing. Keeping the pre-send confirmation step means users retain final control over external communications.
The Battle for AI Entry Points in Apple's Ecosystem
Messaging apps are among the most-opened applications on phones, making them prime territory in the battle for AI 'entry points.' With the Apple Messages plug-in, ChatGPT reaches the most core personal communication scenarios on Apple devices — far deeper than a standalone chatbot app. For OpenAI, this means stronger user stickiness and richer behavioral data scenarios.
At the same time, users face a trade-off: convenience for privacy, or a balance between the two. OpenAI has at least taken a relatively moderate starting stance — local processing, no indexing, and a pre-send confirmation step. How this feature evolves will depend on the interplay between user trust and technical capability.
Sources
Frequently Asked Questions
Q1: What is the ChatGPT Apple Messages plug-in?
A: It is a plug-in OpenAI launched on August 20, 2026 that lets users connect their Apple Messages inbox to ChatGPT, enabling AI to sort, analyze, edit messages, draft and send texts, and search message history.
Q2: Does the plug-in read and upload all my messages?
A: OpenAI says the plug-in runs locally on the user's machine and does not create an index of all messages. However, technical details remain unclear, and it is still to be confirmed which operations trigger cloud processing.
Q3: Why should I be cautious about persistent approval mode?
A: Because with persistent approval enabled, ChatGPT no longer needs your confirmation before each message, removing your final chance to review. If AI misinterprets your intent, an ill-phrased message could be sent, so keeping pre-send confirmation is recommended.
Q4: What use cases does the plug-in support?
A: It supports both personal use (sorting messages, suggesting replies, sending texts) and professional use — the plug-in is compatible with Codex and ChatGPT Work for drafting, analyzing, and editing work messages.
🔧 Related AI Tool Recommendations
Quickly extract core content from long texts to digest messages and documents
Automatically summarize meeting key points and action items
Quickly convert voice messages and recordings to text for easy organization
Conclusion
The ChatGPT Apple Messages plug-in marks AI assistants evolving from 'chatbots in a dialog box' into 'intelligent agents that take over personal communication workflows.' Being able to organize, draft, and even send messages means AI has entered one of the most private and frequent scenarios in the phone experience.
At the same time, the local processing, no-index promise, and pre-send confirmation reflect OpenAI's cautious attempt to balance convenience and privacy. For regular users, whether this feature is worth adopting ultimately depends on how much 'speaking power' you are willing to hand to AI — and either way, keeping a final human review step should be the bottom line.