OpenAI’s August developer updates brought together changes to Codex, ChatGPT workflows, agent interoperability, APIs, image generation, and developer events. In an August 31 roundup, @OpenAIDevs described a month focused on expanding where developers can work, how agents interact with software, and how teams build with OpenAI tools.

Original post on X

Load the post to view it as published on X. X may receive connection data.

View the original post on X ↗

Codex and ChatGPT move across more workflows

The roundup presented Codex as expanding across platforms, browsers, and collaborative workflows. It also highlighted additional browser support, pointing to a broader range of environments for development work.

Several updates focused on context and collaboration. Computer History was described as bringing more of a user’s work context into ChatGPT, with the feature also rolling out in the European Economic Area, the United Kingdom, and Switzerland. The post separately highlighted site building with a team and said ChatGPT Work and Codex could do more across browsers and files.

Taken together, these announcements suggest a shift toward development workflows that connect the tools people use to browse, manage files, collaborate, and build. The supplied roundup does not provide feature-level specifications, so the practical scope of each individual change remains limited to those broad descriptions.

WebMCP and plugins open new paths for agents

WebMCP was another major theme. The announcement described it as a way for agents to interact with apps and websites, then pointed developers toward a challenge built around putting WebMCP to work.

The post also highlighted plugins for compatible agent clients through a shared standard. It framed plugins as a way to extend what developers can do with ChatGPT across areas including messaging and search. This places interoperability alongside model and workflow improvements in the month’s developer coverage.

For developers, the common thread is the ability to connect AI systems with more of the software and services involved in everyday work. The roundup presents that direction, while leaving the implementation details to the individual announcements and developer materials.

API and image-generation updates

The August update included changes to the API as well. GPT-5.6 Sol was described as becoming cheaper and easier to manage as usage scales. The post also said GPT-5.6 appeared in more places where developers build for production.

Image generation received a separate update through GPT-Image-2, which could generate transparent assets for websites, games, and other projects. Transparent output can fit directly into builds that require assets without an opaque background, although the roundup does not provide further technical specifications.

These items broaden the update beyond interfaces and workflow tools. The announcement covered both the models developers use in production and the visual assets they may use in digital products.

DevDay Exchange takes the developer conversation on the road

OpenAI also announced that DevDay Exchange would travel to different locations. The roundup closed by pointing to community builders who had already put the new capabilities to work.

That combination of product updates, developer challenges, and community activity gives the August roundup a wide scope. Rather than focusing on one model release, it connects changes across coding workflows, browser and file access, agent integrations, plugins, APIs, image generation, and developer events.