ChatGPT just released a bunch of new upgrades that make it way more useful for real work. The biggest changes include a new Chrome extension, lightning-fast video analysis, multi-tab context, goal automation, skill recording, improved voice mode, and better ways to connect ChatGPT to the tools you already use.
This is a big shift. ChatGPT is moving beyond being a place where you type a question, get an answer, and move on. It is becoming something that can understand what is happening across your browser, help plan multi-step tasks, remember context, and eventually handle repeatable work on your computer.
1. The New ChatGPT Chrome Extension Changes the Workflow
The new ChatGPT Chrome extension is probably the most exciting update because it brings ChatGPT directly into the browser. Instead of constantly copying information from a webpage, pasting it into a chat, explaining what you need, and then moving back and forth between tabs, you can give ChatGPT the page as context immediately.
That sounds simple, but it is a huge workflow improvement.
For example, imagine you are researching a topic and land on a detailed article. You can add that tab to ChatGPT and ask it to:
- Summarize the page.
- Turn the information into a short-form video script.
- Extract the key takeaways.
- Rewrite the content for a different audience.
- Create a content outline from the research.
- Compare the page with information from other open tabs.
In the past, that meant manually copying text into ChatGPT. Now, the browser tab itself can become the source material. You can ask for a summary and a 40 to 45 second script, for example, without doing the tedious part yourself.
The extension also gives ChatGPT access to more of your workspace, including tabs, files, folders, apps, plugins, chat history, and custom skills. That does not mean every task should be automated blindly. It means you can bring the right context into the conversation much faster.
2. Summarize Long Videos and Podcasts in Seconds
One of the strongest use cases is fast video and podcast analysis. A long podcast can contain great information, but realistically, most people do not have 90 minutes to spare every time they find something interesting.
With the upgraded browser workflow, ChatGPT can analyze a long video or podcast and return a concise summary with timestamps. That means you can quickly see the major ideas, decide what matters, and jump to the exact sections worth exploring further.
This is useful for:
- Researching industry interviews and expert podcasts.
- Pulling ideas from long educational videos.
- Creating notes from meetings or presentations.
- Finding important sections of a discussion quickly.
- Turning lengthy source material into content ideas.
The important part is not just speed. It is context. A useful summary tells you what was discussed, why it matters, and where it happened. Timestamps make the output much more practical because you can verify or revisit the original material without searching through the entire recording.
3. Add Multiple Tabs and Combine Your Research
Being able to mention multiple browser tabs is another massive upgrade. You might have release notes open in one tab, several posts on X in other tabs, an article in another tab, and a video or document somewhere else. Instead of treating every source separately, ChatGPT can combine them.
This is exactly the kind of workflow that makes research dramatically easier. You can collect the latest information from several sources and ask ChatGPT to turn it into one coherent output, such as a report, script, article outline, briefing document, or action plan.
For content creation, this is especially powerful. You can gather source pages, attach the relevant tabs, and use a custom writing skill to create a first draft based on everything you collected.
It also makes multitasking easier. One tab can be handling a larger task while you open another chat and work on something else. You are no longer stuck waiting in one place for every task to finish.
4. Goals Let ChatGPT Pursue an Outcome
One of the most important changes is the ability to set a goal. This is a much more powerful way to use AI than simply issuing one small command at a time.
Instead of saying, “Search for this,” then “Summarize that,” then “Find another source,” you can give ChatGPT an outcome. For example, you might ask it to find and summarize all major ChatGPT changes from the previous seven days.
Once the goal is set, ChatGPT can pursue it, gather relevant information, and return when it has completed the work. You can pause, edit, or change the goal as needed.
This is the mindset shift:
- Basic prompting: Tell AI a single task.
- Goal-based prompting: Tell AI the result you want and let it work through the steps.
That is where AI starts becoming less like a chatbot and more like an assistant. You are focused on the desired outcome rather than micromanaging every action required to get there.
Use Plan Mode Before Taking Action
Plan mode gives you a way to map out a task before ChatGPT carries it out. You describe what you are trying to accomplish, and ChatGPT builds a structured plan instead of immediately taking action.
For instance, if you wanted to research highly speculative meme coin opportunities for a specific month, plan mode could ask questions about your time horizon and risk tolerance before producing a research plan. In the example workflow, the investment horizon was weeks to three months, with a maximum-speculation risk profile.
That does not make speculative investing safe, and it should not be treated as financial advice. What it demonstrates is how ChatGPT can clarify constraints before producing a more relevant plan.
Once the plan is generated, you can expand it, change it, or set it as a goal for ChatGPT to pursue. That is a much smarter workflow than asking a vague question and accepting the first answer.
5. Record Skills to Turn Repetitive Work Into Repeatable Systems
The new Record a Skill feature could be one of the biggest long-term changes. It allows you to record what you do on your computer and turn that process into a reusable skill.
Think of a skill as an SOP for ChatGPT. If you repeatedly follow the same process, such as collecting information, formatting a report, preparing a document, organizing data, or creating a content workflow, you can capture the process instead of rebuilding it every time.
This is incredibly useful because repetitive computer work is exactly where AI agents are heading. The prediction here is simple: within the next six to twelve months, people may not be personally doing many of the repetitive tasks they currently perform at a computer.
The best preparation is to begin documenting and recording your workflows now. Every skill you record gives you a clearer system for the work you already do. It also creates a starting point for future automation.
Start with tasks that are:
- Repeated weekly or daily.
- Based on a clear sequence of steps.
- Time-consuming but not highly creative.
- Easy to verify once completed.
- Built around the same tools, files, or browser actions.
6. Highlight Content and Hand Off Context
The extension also supports highlighting content and handing that context directly to ChatGPT. This is useful when you do not need help with an entire page. Sometimes you only want ChatGPT to rewrite one paragraph, explain a particular section, improve a sentence, or extract an idea from a selected passage.
Highlight the relevant section, provide the instruction, and ChatGPT can work from that precise context. For writers, researchers, marketers, and students, this makes editing and analysis much more direct.
Combined with browser history context, this creates a much more connected experience. ChatGPT can better understand what you have been working on and why a selected piece of information matters.
7. The New ChatGPT Voice Mode Works Across Your Computer
ChatGPT voice mode is also getting a major upgrade. The new experience is designed to work across projects, connected apps, files, documents, and desktop tools.
Voice becomes much more useful when it is not isolated inside one chat window. You can use it to brainstorm, work through a problem, prepare for a meeting, learn a topic through natural back and forth, or keep a project moving while you work.
With screen context enabled, ChatGPT can use active window images and text when you ask about what is on screen. It can also work with connected tools and apps, including platforms such as Slack and Notion.
The biggest takeaway is that ChatGPT voice can now become available wherever you are working. Rather than opening a separate chat every time, you can bring voice assistance into your broader desktop workflow.
There is also a feature referred to as a “pet,” which keeps ChatGPT accessible on screen so you can chat with it from different places on your computer. The name is a little weird, but the idea is useful: ChatGPT everywhere.
8. API Pricing and Faster Model Options
There are updates for API users as well. Starting July 30, the listed API pricing for Terra is $2 per million input tokens and $12 per million output tokens. Luna is listed at $0.20 per million input tokens and $1.20 per million output tokens.
Sol pricing remains unchanged, but a newer version of Sol is available at a higher cost with roughly twice the speed while maintaining its capabilities.
For anyone building AI workflows, token pricing matters because it directly affects the cost of research, automation, content generation, and agent-based tasks at scale. Faster models can be worth the added cost when time savings and responsiveness matter more than the lowest possible price.
9. Settings Power Users Should Turn On
If you want the best ChatGPT experience, do not ignore the settings. The recommended options include:
- Higher intelligence for more capable responses.
- Ultra effort for deeper task execution.
- Dictation for faster voice-based input.
- Cloud browser settings configured for the way you want to work.
- Voice intelligence settings adjusted based on speed versus quality.
- Model selection between live, advanced, and standard options.
If you only want quick, instant responses, choose the settings that prioritize speed. If you want ChatGPT to handle more complex work, research, planning, or computer control, prioritize higher intelligence and stronger reasoning.
You should also explore plugins and MCP-enabled apps. If an app has an MCP server, ChatGPT can potentially connect to it. That expands ChatGPT beyond a standalone tool and lets it work with the software already central to your workflow.
10. Stop Using Basic Prompts
One of the biggest mistakes people make is still using basic prompts. A request like “Create a blog post about AI agents” might produce something decent, but it is not going to produce the best possible result.
That prompt is too broad. It does not explain the audience, desired structure, research depth, style, length, objective, image needs, or final format.
A stronger prompt should include clear context, task details, and output expectations. A prompt optimizer can help turn a vague idea into something much more specific.
For a 1,000-word research-based blog post, a better prompt can define:
- The topic and target audience.
- The desired word count.
- The tone and writing style.
- The research expectations.
- The article structure and headings.
- Header image and inline image requirements.
- Image prompts to generate the visuals.
That is why a tool like MyPromptBuddy is useful. It helps optimize prompts, save your best prompts as shortcuts, and avoid losing workflows that already work well.
When you improve the prompt, you improve the result. Instead of getting only a block of blog text, you can get a complete content package with an organized article, image prompts, and visual assets ready to use.
ChatGPT Is Becoming an Operating Layer for Your Work
The real story behind these updates is not just that ChatGPT can summarize pages faster or speak with you through voice mode. The bigger story is that ChatGPT is becoming integrated into the way work actually happens.
It can understand browser tabs, work with files, connect to apps, help you plan, pursue goals, remember context, and record repeatable skills. That is a completely different level of usefulness from simple question-and-answer prompting.
The people who get the most value from this will not just ask ChatGPT random questions. They will build workflows around it. They will record skills, create prompt templates, use plan mode, connect their apps, and start thinking in terms of outcomes rather than individual tasks.
The move now is simple: start experimenting with the extension, begin recording the repetitive work you do, and stop settling for vague prompts. The more clearly you tell ChatGPT what success looks like, the more powerful it becomes.
FAQ
What can the new ChatGPT Chrome extension do?
The extension can bring browser tabs into ChatGPT as context, summarize webpages and videos, combine information from multiple tabs, highlight selected content, access connected tools, and support goal-based workflows.
What is ChatGPT skill recording?
Skill recording lets you capture a process you perform on your computer and turn it into a reusable skill. It works like creating a standard operating procedure for ChatGPT to repeat later.
Why are detailed prompts better than basic prompts?
Detailed prompts provide context, constraints, goals, format requirements, and output expectations. This helps ChatGPT produce a result that is more relevant, complete, and useful.
How does ChatGPT plan mode work?
Plan mode creates a structured plan for a goal before taking action. It can ask clarifying questions, identify steps, and give you a chance to review or change the plan before turning it into an active goal.
Categories: Artificial Intelligence, ChatGPT, Productivity, AI Tools
Tags: ChatGPT updates, ChatGPT Chrome extension, ChatGPT voice mode, AI agents, record skills, prompt engineering, MyPromptBuddy
Want better results from every ChatGPT session? Try MyPromptBuddy for free, save the prompts that work, and start building a system instead of starting from scratch every time.
