zhangchengzc
zhangchengzc on Hacker News
Posted by
zhangchengzc on Wednesday July 30, 2025 @08:33AM
from the graphvizonline.net dept.
Clean web tool for rendering Graphviz code into PNG, SVG, and other formats. Works great with AI-generated diagrams - just ask Claude or ChatGPT to write the Graphviz code, then paste it in to visualize. No installation needed, runs entirely in the browser.
Really solid deep-dive into context engineering for AI agents from the Manus team. Goes beyond theory with practical implementation examples and real-world use cases. Available in multiple languages too, which is nice for international devs.
Posted by
zhangchengzc on Saturday July 19, 2025 @02:26AM
from the aicodingtools.blog dept.
Posted by
zhangchengzc on Thursday July 17, 2025 @12:50AM
from the aicodingtools.blog dept.
Solid walkthrough of the A2A Development Kit (ADK) with a practical demo showing how to build agents from scratch. Good timing with their recent site upgrade that added tag filtering - makes it easier to find these kinds of hands-on tutorials when you need them.
Posted by
zhangchengzc on Tuesday July 15, 2025 @04:17AM
from the a2aprotocol.ai dept.
Nice update to the A2A Protocol site - they just rolled out tag-based filtering for their dev blog. The Python section has some solid examples, including a brand new expense reimbursement agent walkthrough. Makes it way easier to find the specific examples you're looking for.
Posted by
zhangchengzc on Thursday July 10, 2025 @12:08AM
from the a2aprotocol.ai dept.
Posted by
zhangchengzc on Tuesday July 08, 2025 @09:13AM
from the a2aprotocol.ai dept.
Perfect timing - just as Linux Foundation announced they're hosting A2A Protocol (at Denver's Open Source Summit), their technical blog is packed with real-world examples across TypeScript, Python, Java, and more. Solid resource as A2A pushes to become the standard for agent collaboration.
Posted by
zhangchengzc on Monday July 07, 2025 @10:52PM
from the gemini-cli.xyz dept.
Posted by
zhangchengzc on Friday July 04, 2025 @07:30PM
from the gemini-cli.xyz dept.
Posted by
zhangchengzc on Thursday June 26, 2025 @04:41AM
from the gemini-cli.xyz dept.
I think this can serve as an MCP server
Posted by
zhangchengzc on Thursday May 22, 2025 @09:50AM
from the show-hn dept.
https://veo3.directory/
I've created Veo3.Directory, a platform that aggregates and showcases videos generated by Google's groundbreaking Veo3 video generation model.Veo3 represents a significant leap forward in AI video generation capabilities, producing remarkably realistic and creative content. As these videos have been appearing across the internet, I wanted to build a central repository where people could easily...
Very good suggestions, I am collecting more fonts, if you have any favorites, please feel free to let me know.
Posted by
zhangchengzc on Saturday May 10, 2025 @12:21AM
from the show-hn dept.
https://papyrusfont.com/
I built a web-based tool that lets you create stunning text designs using the iconic Papyrus font. Features include:- Real-time text customization (color, size, background)
- Advanced effects (shadows, outlines)
- PNG export with transparency support
- Mobile-friendly interface
- Available in English and ChineseBuilt with Next.js, React, and Tailwind CSS.
Posted by
zhangchengzc on Monday April 28, 2025 @09:33AM
from the show-hn dept.
https://iammusicfont.com/
I built this tool to let anyone create text in the distinctive layered typography style of Playboi Carti's 2025 "MUSIC" album cover. It combines a bold background font with a stretched foreground layer to create that unique aesthetic. Just type your text, customize if needed, and download as PNG. Works on mobile too. No account required, completely free to use.
Posted by
zhangchengzc on Friday April 18, 2025 @10:56PM
from the show-hn dept.
https://qwq32.com/
I've built an aggregator for free Large Language Models that provides answer generation services. The project aims to make powerful AI accessible to everyone as I believe free LLMs may become a significant trend.Currently supported models:Qwen Series
- Qwen 32B: Alibaba's 32B parameter model for Chinese/English content
- Qwen 2.5 7B Instruct: Lightweight, responsive daily assistantDeepSeek...
Posted by
zhangchengzc on Sunday April 13, 2025 @10:53PM
from the github.com dept.
I have a question: with the agent.json in A2A being so complex, how do LLMs call Agents?
Most discussions suggest that A2A and MCP have a complementary relationship. But I don't think so. If they were complementary, then MCP would only handle local function calls, with all remote services being implemented through A2A. However, from my perspective, A2A can completely handle everything.
I have tried it out, and the more modern design is very convenient.
I believe this is a more powerful and versatile approach than MCP.
AI systems like Manus that can directly operate browsers make APIs not even necessary.