We were chuffed to have been chosen to host a talk on The Power of Product Screenshots in your Help Documentation at the Write the Docs Atlantic conference on September 12. It was a great couple of days learning all things software documentation with the passionate and engaging community that is Write the Docs. Thank you to everyone who took the time to listen to our talk and for all the thoughtful questions throughout.
Q: Do you think that in the foreseeable future AI image recognition could help (maybe with ALT text) in a "bulk" automated screenshots pipeline? - Federico Marafin
TLDR; Yes! (*you nailed it) :-)
So much so that you pasted part of our roadmap in your question. This is exactly what we ultimately want to add to the very pipeline we are in the midst of building. First, put in place a pipeline that can spit out 10,000 PNG’s every night, next add metadata, such as valuable ALT text to all of these for additional value.
Or more generally, this is exactly what I think anybody would need to do in the future, even if manually extracted on a per screenshot basis using a small command line tool or script like Ian suggested, etc.
Ian: There are some tools around like Selenium: browserstack.com/guide/take-screenshots-in-selenium
You can see a summary of the talk and find a link to the recording here.
And you can also find a link to the other great questions from the chat below:
- Recommendations for writing good Alt Text for your screenshots
- Does the title attribute on images help with SEO?
- When are product screenshots most helpful?
- When to add images in developer-focused documentation?
- Recommendations for image file formats
- What is the right ratio of images to text?
- Are GIFs better than screenshots with callouts?
- How to automate screenshots that illustrate product capabilities?