ฝัง
- เผยแพร่เมื่อ 6 ก.พ. 2025
- Today, we are announcing AI Attachments in Raycast. You can now add text files, pdfs and images to the AI Chat as context, which makes tons of workflows a lot easier. Check out this tutorial for tips on how to use it.
If you're not a Pro subscriber yet, start a 14-day free trial → ray.so/free-trial
Download Raycast → raycast.com
Follow Raycast → / raycastapp
I've been a Raycast pro user for about a year now, and it just keeps getting better. I use it every day and honestly don't think I could live without it going forward. Thank you for all the work you've done on this product. I'm excited to see what features y’all come up with next.
One of the good uses for vision is to get it to describe the image for you to add into the accessibility descriptors for low vision for your presentations and documents that you share.
That sounds useful!
Finally you have convinced me to subscribe to Raycast Pro.
Welcome!
@@raycastapp If I have you here right now anyway. Maybe you can answer a question I've had about Raycast AI for a while?
Are you planning to add something like Raycast AI Agent? That could perform system tasks, for example, by using Raycast extensions for the user?
Open Raycast, type a goal, Raycast performs all tasks to reach the goal as long as it's within Raycast's capabilities?
Your videos are simple, not overwhelming, and valuable. They happily pushed me to become a Pro user, and I'll never return. Please keep a regular cadence of these tips and tricks. You have plenty to be proud of with Raycast, and you keep your subscribers happy. Thank you.
On the file attachments, I travel extensively for work and use it to process messy hotel receipts. When doing expense reports, I have to break down and total all the categories - room rate, as many as six different taxes per day, meals, laundry, parking, etc. I can use AI Chat with Raycast to scan through all of them and give me grand totals for each category that I can then transfer to an expense report. Previously, I had to make a spreadsheet of all the breakdowns. Now I can do it all in seconds with Raycast Pro. Amazing.
The new AI functionality in Raycast is simply fantastic! I can't imagine living without Raycast Pro anymore. The integration with ChatGPT and other AI tools makes my productivity skyrocket. Being able to ask questions, generate content, and automate tasks directly within Raycast is an incredible experience. Congratulations to the Raycast team on this excellent update! The investment in the Pro plan is definitely worth it. 🙏🏽❤🚀
Please keep doing what your doing with this app! It is the best thing to happen to the Mac! You guys are the absolute best and I couldn't imagine my workday without Raycast w/ Pro. 🤯
I’m a Front End developer and I gave AI Chat screenshots of a UI I was building and asked it for feedback on it. It gave me specific examples of animations and interactivity I could add that were genuinely good ideas! This feature is a game changer and solidifies my decision to make Raycast the only AI tool I pay for.
That's interesting 🤔
When reading technical blog posts that explain the architecture of an application or other techy concepts, I often find it useful to convert accompanying diagrams into Mermaid format. This allows me to modify and adjust the diagrams to suit my project needs.
I have been waiting for this feature since I signed up for PRO. It helps a lot with my coding school projects. I can simply input the PDF of the project explanation and get some help from the model.
Nice use case!
Adding images is excellent! I always include text alternatives to my images that I post online and adding an image from my clipboard and sending it to an AI preset with the best practices for alt-text would save me a lot of time
Using vision on a designer/dev standpoint means you can feed it web components and get it to write code back, might not be perfect but gets a head start, another is for re prompting, get the ai to generate a prompt of an image you like so you can add it back to mid journey or other image generation models.
The most awaited feature. Thank you for this!
I think AI Vision can be used for more creative tasks, such as creating a detailed floorplan for a house with the visual references you previously attached; or give a sketch of a logo and having AI give me recommendations on how to perfection it
Excellent video guys! Going Pro soon!
Well I'm convinced now. Getting pro today. One last we need is a VSCode extension to directly ask questions about my code without manually sharing my files.
We’re working on allowing extensions to leverage AI attachments really soon!
Yeah this would be really cool
Vision can be used for translation. Or the screenshot example you shared. Extract text items and organize them etc
I would use the vision feature to create descriptions and highlights for products in e-commerce store, or even the hashtags and descriptions to share it in social media. It would be amazing to have a shortcut for that and be able to save the conversation in a txt or any text file.
Ok, my use case is learning a new software, like Excel. I ask AI for a breakdown of a function or capability of the software. I attach a small screenshot of some of the data I’m working with and ask AI to come up with an example of how to use the function in the context of my data. Then I’ll try implement it a couple of times. If I run into errors, I can screenshot my excel workbook (formula bar, output, range of cells - everything!), paste it in, and simply ask something like: why am I getting an error? So useful! I can turn AI into my ultimate teacher: patient, enthusiastic, aware of my context.
That’s actually very clever
@Raycast Suggestion:
I would love to be able to increase the size of the text and window in the AI chat. I find it hard to use on larger monitors as the text is small. Incorporating cmd + Plus or Minus like other apps do would be a quick fix if possible
Yesss!!! This is amazing!
Will it be possible to add attachments to AI presets as context/knowledge base?
That's something we talked about, I personally think it's a really good idea! But let's see how it evolves.
first off u are a god
vision feature will help me iterate product ui
Please allow to attach multiple files / clipboard at once, by enabling multiple selects from clipboard / screenshot history, that would improve a lot the UX
Will try OCR of my handwritten notes and next use case is scanning my receipts with item categories
Vision would be really useful to evaluate design and UX flows.
I'm using your pro subscription plan, and I'd be even more grateful if you could add two features that are important to me.
1. Export generated tables from AI chats
2. Formatting formulas in the same place 😊
In any case, I am very grateful to you for your work, for your support and its development. Your product increases my productivity hundreds of times, which allows me as a Project Manager to work at super-high-speed.
I wanted to generate an image with AI of my building's lobby. However I cannot describe it with enough detail to get a 1:1 render. It would be great if I could take a photo with my phone and upload it to AI, then have it add or remove things. This would be really helpful to present things to the condo board.
For vision use-case: I need to extract data from the image and put it in correct json format. This is what I'm using claude sonnet for now. But I'm interested if Raycast can help me with that.
If there will be an extension for VSCode, please don't forget an extension for JetBrains IDE.
Excellent 🎊I have been hitting the rate limit recently 😅this will make it even more frequent 🤣
Love to see it!
I often prefer to share screenshots with AI instead of plane text or code just because it’s easier with shortcuts (muscle memory).
Fair enough!
Thanks for this🎉
Hey could you implement the inline LaTeX support for the AI Chat and even quick AI? I would love that :)
Watch out for the next release!
Really cool. What about Data Privacy? Are the attachments used for training the LLM?
Raycast just transfers your input to the AI models, so everything you share is used by the models. I think you should read model privacy policies that you are interested in.
If this works without bugs, this is what tips the scale for me to become a paying user
You can now start a 14-day free trial to test it out yourself 😄 Link in the description
I use it mostly to generate boiler code for UI passed as image.
Nice! Does it work well? At least as a starting point
For smaller components it does well!
why we cannot attach .xlsx file on AI chat?
Excellent, I don’t have good front end skills then I will use it with my front end presets for making prototypes faster copying and pasting designs 😅🤦♂️
Cool!!!
It can give a file after making changes too?
Not yet. Currently it only accepts files as input
@@raycastapp it will be cool too ❤️
cool! it would be awesome if we also had raycast on ios. that'd give me a solid reason to go pro.
Soon!
how about "convert image into SVG" using vision
Hmm, I seem to be unable to attach coding files such as .ts or .tsx. They're greyed out and are not accepted when dropping in. Model picked was GPT-4o. Am I doing something wrong?
Amazing release btw!
Oh, thanks for pointing that out, we’re investigating. Could you share some details with us at feedback@raycast.com? Some of the team could reproduce it with .ts files, but .tsx work fine. We’re working on it 🤛
Are you at all concerned that, if I ask Raycast AI to read a legal document like a lease-agreement and it hallucinates, Raycast may be found liable for damages?
We really need an iOS app of Raycast to use my AI subscription on my phone ❤
Coming soon!
@@raycastapp Crazy!! Have been waiting this for soooooo long! Really looking forward to it :)
I hope they don't charge extra bucks for using the iOS app 😅
Would you nice to use my own API key so that I can use GPT4 without paying double.
Is Raycast able to rename files for me? I want to drop a bunch of files into AI chat and explain the format I want (which is based in file contents). This would save me so much time…
Raycast AI can’t currently execute functions like that at the moment, but an extension could leverage the AI API for that!
Please add Word documents!
5:37 you guys add this so I can unsubscribe chatgpt