Developer Experience @OpenAI | ex-Slack SDK & DevRel, SmartNews VPoE | ja-jp 🇯🇵👉 @seratch_ja
Tokyo, Japan
Joined December 2010
- Tweets2.9K
- Following230
- Followers1.7K
- Likes1.1K
Kazuhiro Sera retweeted
Turns out building with @ChatGPT Sites is pretty addicting. So is clicking cookies. 🍪
I made a little demo that does both, with a database and an MCP server so ChatGPT and Codex can join in.
Come hang out. I hope it gives you an idea for something fun to build.
Chapters:
00:00 A cookie clicker for humans and AI
00:27 The prompt I started with
00:41 Saving clicks in a database
01:16 Adding MCP tools
01:29 MCP in plain language
02:09 Testing with MCP Inspector
03:11 Asking ChatGPT for the click count
03:40 Clicking a cookie with Codex CLI
04:10 What will you build?
Kazuhiro Sera retweeted
Had a very fun panel chat with @coreyching and Bonnie 😁 The crowd was amazing!
#OpenAIDevDayExchange
Kazuhiro Sera retweeted
Great to be in Tokyo with @gabrielchua and @seratch for OpenAI DevDay Exchange 🇯🇵
Excited to share what we’ve been building with the developer community here!
よろしくお願いします！
Kazuhiro Sera retweeted
Valibot made it into the OpenAI Agents SDK. 🔥
Thank you for implementing the Standard JSON Schema spec!
OpenAI Agents SDK for TS now supports Standard Schema.
This means you can use not only Zod, but also any other schema library compatible with the spec, such as Valibot.
See openai.github.io/openai-agen… for details! 👋
Kazuhiro Sera retweeted
To celebrate the realtime model release, @seratch has updated RealtimeAgent to use gpt-realtime-2 by default in the Agents SDK
📷 (WebRTC & WebSocket): openai.github.io/openai-agen…
📷 (WebSocket): openai.github.io/openai-agen…
Kazuhiro Sera retweeted
If you want to try it out Sandbox Agents are now also in the Agents SDK for TypeScript!
Thanks @seratch 🙌
github.com/openai/openai-age…
Kazuhiro Sera retweeted
Excited about the Agents SDK updates we just launched.
Check out my cookbook on using it with sandboxes for code migration:
developers.openai.com/cookbo…
Kazuhiro Sera retweeted
Build long-running agents with more control over agent execution.
New capabilities in the Agents SDK:
• Run agents in controlled sandboxes
• Inspect and customize the open-source harness
• Control when memories are created and where they’re stored
Kazuhiro Sera retweeted
Today, we’re launching a Rather Large™ update to the OpenAI Agents SDK.
Agents SDK now allows you to scale Codex-style agents in production, without building the whole harness yourself. We’ve brought all of the stuff of modern agents: computer-use, skills, memory, compaction, and more to the same platforms you’re already using.
Kazuhiro Sera retweeted
We use skills to maintain our Agents SDK repositories through repeatable workflows for verification, integration tests, release checks, and PR handoff.
Here’s how it works:
developers.openai.com/blog/s…
New post on the OpenAI Developer Blog: how we use skills for open-source maintenance, from planning and coding to testing and release-readiness checks with GitHub Actions. Hope it's useful for your projects too 🙌 developers.openai.com/blog/s…
Kazuhiro Sera retweeted
I became a developer because it allowed me to turn my ideas into reality...
I joined Twilio because APIs allowed me to build more and have my code engage with the world...
Grateful to be part of bringing Codex to the world to empower everyone to build. It's been amazing to see friends and family like my brother build and ship apps, something they weren't able to do before.
We've reached a world where you can just build things! ❤️
Kazuhiro Sera retweeted
2025 wasn’t about one launch - it was the year of stronger models, agent-native APIs and SoTA tooling!
We shipped:
- reasoning that converged (o1 → o3/o4-mini → GPT-5.2)
- codex as a coding surface (GPT-5.2-Codex + CLI + web/IDE)
- real multimodality (audio + realtime, images, video, PDFs)
- agent-native building blocks (Responses API, Agents SDK, MCP)
- open weight models (gpt-oss, gpt-oss-safeguard)
And the curve moved fast (4o -> 5.2):
GPQA 56.1% → 92.4%
AIME 9.3% → 100% (!!)
SWE-bench Verified 33.2 → 80.0 (!!!)
Full recap:
developers.openai.com/blog/o…
Kazuhiro Sera retweeted
The legend continues!
Shohei Ohtani is the NL MVP for the second straight season!
ALT Shohei Ohtani NL MVP Pictured: Shohei Ohtani smiles while wearing a road gray Dodgers uniform. Below, Ohtani yells while wearing a home white Dodgers uniform. In the lower right corner is Ohtani’s signature in Japanese.
Kazuhiro Sera retweeted
This week we held our biggest DevDay yet. And honestly we couldn't have done it without Codex!
I went ahead and collected a couple of examples of how Codex helped us across the entire show
developers.openai.com/blog/c…
Kazuhiro Sera retweeted
GPT-5 is astonishingly good at frontend development.
Here’s a gallery of dozens of beautiful demo applications, each generated entirely in a single GPT-5 prompt, without writing any code by hand! ✨
Kazuhiro Sera retweeted
I've been mind blown by some of the apps that GPT-5 has been creating entirely on its own. @seratch and @kagigz have been collecting some incredible examples in the docs 👇
platform.openai.com/docs/gui…
Kazuhiro Sera retweeted
Heads-up for developers trying gpt-oss: performance and correctness can vary a bit across providers and runtimes right now due to implementation differences.
We’re working with inference providers to make sure gpt-oss performs at its best everywhere, and we’d love your feedback!