I asked @taskletai to generate a custom font from just a single picture of my handwriting
two takeaways:
- the models are getting good at random tasks
- my handwriting is bad enough I should have become a doctor
The best models from Anthropic, OpenAI, and Google - all in one place.
Dial the effort up or down depending on the task. And when a top lab releases a better model, your agents can use it in Tasklet right away. No migrations, no waiting.
We're testing others, including open source models - which one do you want to see next? ✍️
Building an automation in Zapier or n8n means wiring up nodes, mapping fields, and debugging a flowchart before anything runs.
That's why @TaskletAI is here: describe the task in plain English and let a long-lived agent figure out the steps, tools, and integrations itself.
Spilling some beans 🤫 This very startup is actually looking for their first design engineer:
https://t.co/0wCaTeN0Av
It's an incredible opportunity, owning the product design end-to-end. They've identified that great design is essential for them.
Funnily, I've told the founder to *not* look for a "design engineer." I told him, the most cracked product designers are all AI-pilled and now working with code anyway. If that's you and you're remotely curious, highly rec to get in touch with the founder.
Oh and, did I mention, it's the Firebase founder?!
Tasklet now supports audio transcription + video generation.
Upload calls, podcasts, meetings, or voice notes → get transcripts, summaries, action items, clips, drafts, and more.
And with triggers, it can run automatically.
Subscribe to a podcast RSS feed, and Tasklet can transcribe every new episode, monitor for topics, summarize the best ideas, and send you the takeaways.
New media comes in. Tasklet gets to work.
https://t.co/CK7D0vyS5d
We're committed to bringing you the best models from every lab.
Starting today you can use GPT 5.5 in Tasklet, and you can customize thinking effort & other settings.
We also now have built-in image generation as well!
Fun fact: the banner image at the beginning of this thread was also generated in Tasklet. ✓
GPT-5.5 and GPT Image 2.0 are available now.
Try them in Tasklet today: https://t.co/us0XyW6kRm
Also live today: image generation in Tasklet, powered by GPT Image 2.0.
The best AI tools become more useful when they’re connected to the rest of your work.
Now agents can generate visuals for campaign briefs, presentations, reports, social posts, product concepts, and more - without sending you to another tool.
Work shouldn't depend on which tab you're in.
Most teams don’t need another dashboard. They need the tools they already rely on to actually work together.
That’s what Tasklet is built for.
Your tools, finally working together. ✓
Stop saving X articles you'll never read.
Setup an agent that you can text/Telegram and that has browser use to scrape and summarize the article.
Here's a simple setup with @TaskletAI. Prompt 👇
Every Tasklet agent now has a built-in browser.
No VM. No setup. No extra cost. Just a fast Chrome instance your agent controls like a human — navigate, click, fill forms, extract data, download files, take screenshots.
If you can do it, your agent can too. 🧵
A few things to try:
🔹 "Screenshot my Google Ads dashboard"
🔹 "Download receipts from Gmail and upload them to my expense tracker"
🔹 "Log into Shopify and find all unshipped orders from this week"
🔹 "Publish the draft blog posts in WordPress"