Skip to main content

Tools worth trying


If you have a good tool, please share it with me πŸ₯°

AI Product​

  • iconifyai: This is a product for AI-generated app icons. The price is around $10 for 15 icons and $15 for 30 icons. From the demos, the quality seems much better than hiring someone on a platform like Fiverr (I'm not sure how it actually is - I'm planning to try it out later). It also seems pretty affordable - probably around $15-20 to get it done properly including trial and error.
  • Magic: This startup recently raised $23 million, and their main focus is using AI to write code. As I said in my Xiaoshengnal newsletter, I think AI is still weak in true creativity. It can probably handle writing interfaces for now, but writing logic is likely too difficult currently.
  • Galileo AI: An AI product for generating UI. The demo looks promising, but I'm not sure how it actually performs. Interested friends can apply to try it out.
  • rationale: An AI assisted decision-making tool, skilled at pros and cons analysis and SWOT analysis. Personally, I feel AI is more comprehensive than humans at searching and listing possibilities, but I'd be a little hesitant to let AI make the final decision. However, I'm more supportive of AI assisting in decision-making.
  • Artifact: This news app is the latest startup from Instagram's co-founders and ex-CEOs. It highlights AI capabilities and has been described by tech media like The Verge as "a kind of TikTok for text".
  • Read Pilot:This is an AI article summarization app developed by an indie dev. I'm recommending it because unlike typical AI summarizers, it summarizes by answering key questions, which I think is a great design. It constraints the scope of the summary, making it more accurate, and readers can quickly judge if the article is worth reading from these summaries.
  • DeepL Copywriting Assistance:I've been using DeepL as a translation tool, and I think the results are even better than Google's. They recently launched a copywriting assistance feature. I tried it out and feel it may be able to replace Grammarly?
  • Eightify: This is a browser extension that uses AI to generate video transcript summaries. For now it only works on Youtube, but I tried it on product review videos and the summaries were pretty accurate - you can directly read the text to understand the pros/cons etc. Very efficient.
  • Queryable: This is an app developed by a Chinese dev that lets you search local photos by text, and it works offline (the models run locally), which is great. It feels like the kind of photo search capability the iPhone should have.
  • D-ID: After AI image generation capabilities went mainstream, I had a feeling video would be next. Runway tried a difficult path, while D-ID is going down a vertical path focusing on realistic human faces. You can take an AI generated character image and import it into this product, enter the text you want them to say, and D-ID can output a voice-over video. Pretty amazing. Though there are some requirements for the input images - some images make the mouth look very distorted! πŸ˜‚
  • Drawthing: I recently had an immature idea - could Mac overtake Windows in the AI era? The reason is much of today's AI computation happens online, with lower privacy protection. In the future, could local AI products have a chance? Based on this thinking, Mac's GPU currently has stronger AI computation abilities compared to Windows - so could it enable an overtake? Similar to how Microsoft wants to overtake Google through AI + search. So recently I've also been trying out some local AI products. The image generation speed is really slow, but the benefit is it's free! Many ChatGPT based products have shut down now because the costs are too high. In the short term, local products sacrifice performance for low cost; in the long run, they sacrifice optimization for privacy.
  • hairstyleai: The use case for this product is also pretty interesting. It uses AI to quickly generate photos of you with different hairstyles. Though you can also just use Midjourney for this.
  • CodaAI: Coda also recently released some AI capabilities, but I haven't gotten access to try it yet. From the demo videos, the overall interaction seems similar to Notion AI, but Notion AI outputs text, while Coda supports generating tables from text input, and it looks like it may even automatically import external data into the tables, e.g. enter "Help me plan an itinerary for Paris" and the generated table includes images of the attractions.
  • Pebblely: This product is probably one of the more reliable commercial applications of AI generated images. Upload product images, describe the desired background in text, and it can generate polished product images.
  • FlowGPT: : A site aggregating ChatGPT prompts - you can find many interesting ideas to try. Frankly, these prompts should be tightly integrated with ChatGPT itself.
  • ChatDoc: A Chinese ChatGPT document reading assistant. Upload PDFs/Word docs and chat to instantly get summaries, key points, details, data analysis, content proofreading... You can also select tables/text and ask targeted questions to get more accurate answers. ChatDoc accurately identifies and parses tables/paragraphs in PDFs so AI answers can be traced back to document content. Support for images, scans, and cross-doc Q&A coming soon.

Prompt IDEs​


  • Chat-Web: A ChatGPT demo web page built with Express and Vue3.

Midjourney Keyword Database​

  • Kalos has collected the names of over 1000 artists and the images generated with different keywords in Midjourney, along with search links for them.