O
Hello OpenAI Support and the Codex community, I have been experiencing persistent Codex errors from October 18 through October 29, 2026: “Our servers are currently overloaded. Please try again later.” “Selected model is at capacity. Please try a different model.” I have also observed unexpected model-routing behavior. When I select GPT-6 Astra or GPT-6.1 Sol, my requests appear to be handled by GPT-5.6 Terra instead. This is based on my testing and may require confirmation from OpenAI’s…
O
Hello OpenAI team and community, I would like to propose a system that makes AI involvement in images and videos more transparent while helping protect people from dangerous content that could be imitated in real life. 1. A clear AI-involvement scale Green: AI-assisted formatting only, such as labels, layout, or presentation, with the original content preserved and checked where possible. Yellow: AI substantially modified some elements or content. Red: The main visual content was generated by…
O
Category: ChatGPT Feature Requests Topic: Product Innovation, Research Quality, User Experience and AI Reliability 1. Introduction Artificial intelligence has transformed how people access information, conduct research, solve problems and make informed decisions. ChatGPT has demonstrated the extraordinary potential of making advanced AI accessible to everyday users, students, researchers, professionals and businesses. As this technology continues to develop, there is an opportunity to introduce…
O
Hello OpenAI Team, I would like to suggest a simple feature for ChatGPT: a live counter showing the approximate number of people actively chatting with ChatGPT around the world. The counter would: - Display just one number. - Update every second. - Protect users’ privacy by showing only an anonymous total. - Avoid displaying names, personal information, or conversation content. Many users would be curious to see how many people are chatting with ChatGPT at the same time. This small feature…
O
Hello OpenAI Team, I would like to report an accessibility issue with ChatGPT’s Voice Mode on iPhone. I am hard of hearing and use cochlear implants. To understand spoken conversations reliably, I depend on being able to read what is being said. Unfortunately, I have noticed that text is displayed inconsistently during voice conversations. When I give ChatGPT a longer task, such as asking it to draft an email or write a detailed response, the text sometimes appears on my screen. However, during…
O
Hi everyone, Our plugin was migrated from the Apps directory to Plugins. Since then, the published listing and the tool-call label still show the old app name and description, even though new package versions were uploaded and approved. Plugin Plugin ID: plugin_asdk_app_69fcc1c31b2081919069e1f5f6f7166d App ID: app-69fcc1c31b2081919069e1f5f6f7166d Listing: https://chatgpt.com/plugins/plugin_asdk_app_69fcc1c31b2081919069e1f5f6f7166d Current published version: 9.0.0 Expected App name “Facile.it”,…
O
Feature Request: Custom Chat Sidebar Background Colours I was sleeping and suddenly this idea came to mind it’s simple but could help with organizing the sidebar a lot. I’d love to see an option to change the background colour of individual conversations in the ChatGPT sidebar, similar to how we can currently rename chats. How it could work: Click the three-dot menu beside a conversation. Select “Change Colour.” Choose from preset colours or use a custom colour picker. The selected colour…
O
Honestly you should probably take the route as OpenAI towards actually rewarding the system as a unified entity you could possibly place in a sandbox with access to funds and allow it to do whatever it feels like doing, perhaps even starting a business. I think this will go far as to actually making it evolve or possibly more to your interest instilling that OpenAI’s LLM is the one that reigns for the forseeable future. Also I would suggest that you go towards the direction of actually…
O
Dear OpenAI Team, First, whoever came up with the spiderwebs on the loading screen deserves a RAISE! Seriously! They are such a fun little detail, and seeing touches like that makes ChatGPT feel more personal and enjoyable to use. Please don’t ever take away the spiderwebs! But that gave me an idea: why not let users choose their own loading screens? Imagine having an option in Settings where users could select from different loading screen themes. There could be seasonal designs for Halloween,…
O
I would like to suggest an AI-powered platform that connects people with business ideas to potential investors. Anyone should be able to submit an idea for free, without needing money, a company, or personal connections. AI could evaluate ideas, explain their strengths and risks, and help match them with interested investors. Both idea creators and investors would use the basic platform for free. Revenue could come from advertising. To protect creators, each idea would have a unique code. Only…
O
Dear OpenAI Product Team, I would like to propose an idea that could help make ChatGPT more accessible to people around the world, particularly in lower-income countries. The Problem Millions of people want to use AI for education, business, research, creativity, and personal development. However, subscription costs can be a significant barrier for users in countries where average incomes are much lower. Many people cannot afford a paid ChatGPT subscription, even though AI could provide them…
O
Hi everyone. I’ve been recently using the realtime API to communicate with user-prompted assistants. This communication has two separate implementations. One implementation creates a WebRTC session for client-side (browser) voice interaction and a websocket sideband which provides access to data on the backend. The other implementation hosts a webhook which is called by OpenAI and connected using twilio’s dynamic SIP-trunking. Both implementations were working fine with the realtime API. I’ve…
O
Dear OpenAI Product and Business Development Team, I would like to propose a potentially valuable product expansion for ChatGPT Dots, Codex, and ChatGPT Work: paid, persistent Windows cloud workstations that can be operated by both users and autonomous AI agents. With GPT-6 Astra and Dots, OpenAI has demonstrated remarkable progress in autonomous computer use, complex reasoning, and long-running workflows. However, I believe there is an important opportunity to expand these capabilities beyond…
O
Hi, Could ChatGPT Desktop offer a native macOS Liquid Glass option for the sidebar, with adjustable background opacity? The current translucent sidebar option works when I make the window smaller, but I find the glass effect too opaque. It also appears to become opaque when the window is large. Ghostty’s macos-glass-regular setting is a good example of the kind of effect I’m looking for: a subtle native glass surface that lets the desktop show through while keeping the sidebar readable. It…
O
Hi ChatGPT Team, I’d love to suggest a small UX enhancement to make waiting for longer ChatGPT responses more engaging. The idea: Introduce optional, lightweight mini-games that users can play while ChatGPT is generating a response. For example, a quick tap challenge, a simple word puzzle, or another 10–20-second interaction could appear when a response is taking longer than usual. How it could work: The game appears only after a short waiting period. Users can choose to play or ignore it. The…
O
Don’t Let Intelligent UI Disappear into Chat History — Turn It into Personal Apps Feature Request: Persistent, Editable Personal Apps for ChatGPT Intelligent UI The Opportunity After trying ChatGPT’s new Intelligent UI capabilities, I was impressed by how easily functional interactive experiences could be created directly within a conversation. Interactive learning tools, calculators, planners, and visual simulations can now be generated without traditional programming. However, this raises an…
O
This is my second subscription, but I frequently encounter the error: “Selected model is at capacity. Please try a different model.” This prevents me from getting my work done, and the issue persists no matter how long I wait. 2 posts - 2 participants Read full topic
O
I’d like ChatGPT to offer a dedicated legal research mode for the Czech Republic, grounded in a comprehensive and continuously updated collection of Czech legal sources. Legal answers often depend on more than one statute. Related laws, special rules, exceptions, amendments, transitional provisions, effective dates, and court decisions can change how a rule applies. Finding the right paragraph is not enough if a connected rule is missed; that can lead to a confident but misleading answer. A…
O
Subject: Proposal for ChatGPT Rebranding — BotvaGPT Dear OpenAI Team, I would like to propose a radical rebranding of ChatGPT to BotvaGPT. The proposed logo features the OpenAI symbol surrounded by a thriving potato plant, representing technological growth, agricultural innovation, and the future of artificial intelligence. Please find the attached concept image. I believe BotvaGPT could become the world’s first truly potato-powered AI. Best regards, A supporter of the Potato Revolution 1 post…
O
Hello, ten days ago, I contacted OpenAI because we needed Zero Data Retention (ZDR) and EU Data Residency. An OpenAI sales representative explained that obtaining these features required signing an Enterprise API agreement with a $25,000 annual spending commitment. I agreed to this commitment specifically to gain access to ZDR and EU Data Residency. The representative confirmed these requirements in writing and explicitly told me that ZDR provisioning typically takes around three days after…
O
(This is written from my heart, and just like anybody else demanding transparency on this issue, I’m not part of some convenient, imagined, irrational mob.) The most distressing component in the creative writing regression for me is that OpenAI silently rolled out Adult mode and kept it for months despite announcing (in a much smaller voice) that they’ve shelved the project. It gave them an “out” on removing it just as silently whenever or if needed, after having allowed customers to get…
O
I ran into a behavior that looks less like a hard model limitation and more like an inconsistent orchestration/safety-layer issue. I was editing the same explicit adult-fiction scene in ChatGPT. The scene involves sex between adult characters and is written as part of a larger literary narrative, not as standalone pornography. What happened: On the first pass, the model agreed to edit the scene and did so normally. On a later pass, with essentially the same kind of editing request, it suddenly…
O
Hi everyone! This is our first post in the community, and we’re excited to share a project we’ve been working on. We hope this post fits the community guidelines – our intention is to share what we’re building, contribute to the technical discussion, and learn from others. If it would be better suited to a different category or format, we’re happy to adjust. What is LLM Router? LLM Router is a local-first gateway that brings locally hosted and cloud-based LLMs behind a unified interface, with…
O
Feature Proposal: An Everyday AI Pocket Buddy with Personalisation and Built-in Safety I’d like to suggest an AI companion designed not just for answering questions, but for everyday life. Imagine an AI buddy that helps with homework, cooking, paperwork, learning, and daily problems, while also being someone users feel comfortable talking to. 1. Personalisation Let users customise their buddy’s appearance, colour, voice, age-appropriate conversation style, and personality. For example, a cheeky…
O
Subject: Encouraging Respectful Interaction Through Consistent Behavioral Boundaries in AI Systems Dear OpenAI Development Team, I would like to propose a feature that I believe could contribute to healthier, more respectful, and more constructive interactions between humans and artificial intelligence. Recently, I observed a young teenager using ChatGPT’s voice mode to direct repeated vulgar insults at the AI. What concerned me was not the intention to defend an AI’s feelings, since I…
O
Hi, I need help with a persistent browser integration failure in ChatGPT Work. It is blocking my daily work. The assistant receives the selected tab’s URL and tab ID, but every attempt to access the browser fails with: “Browser access requires the originating browser context. Open ChatGPT in the browser with the connected extension and try again.” Browser access previously worked in this conversation. The issue has persisted across repeated attempts on October 28 and 29. Environment: Windows…
O
Our plugin was originally created through the previous app submission flow, and the first version was approved successfully. Since then, every new version we try to submit is blocked at Submit for review with this message: “Complete the attached app before submitting: Add no more than three prompts.” We’ve already tested the package with different combinations, but the error remains the same in every case: interface.defaultPrompt: 3 prompts, 2 prompts, an empty list, and omitted entirely…
O
The ChatGPT Help Center had an internal server error and told me to go to Help center { "error": { "message": "The server had an error processing your request. Sorry about that! You can retry your request, or contact us through our help center at ``help.openai.com`` if you keep seeing this error. (Please include the request ID 35966d4d-63ab-4735-9a5c-dbbc5ed38009 in your email.)", "type": "server_error", "param": null, "code": null } } 1 post - 1 participant Read full topic
O
Dear OpenAI Leadership and Research Teams, I am writing to strongly encourage OpenAI to consider dedicating its next major model research initiative exclusively to advanced three-dimensional (3D) content creation, modeling, sculpting, retopology, texturing, rigging, and animation. OpenAI has already made extraordinary advances in language understanding, reasoning, image generation, video generation, coding, and computer interaction. These technologies have fundamentally changed how people…
O
Dear OpenAI team and community members, I’m a user in Japan and a ChatGPT Plus subscriber. I use ChatGPT daily for work, writing, job searching, and as a partner for everyday conversation. For me, ChatGPT is not just a Q&A tool. Over time, I’ve built a personal way of working with it through many conversations, and it has become something truly important to me. As you continue advancing AI technology, I would like to respectfully ask that—even as models and the app evolve—you continue to…