
by OpenAI
OpenAI Codex is a software engineering agent that writes code, fixes bugs and runs tests on tasks you give it. It is available in the ChatGPT web app, a command-line tool and IDE extensions.
AgentSubscription
Step 1: Open Codex in the ChatGPT web app, install the CLI, or add the Visual Studio Code extension.
Step 2: Connect the repository you want the agent to work on.
Step 3: Describe the task, such as a feature to write or a bug to fix.
Step 4: Let Codex work in its environment, reading and editing files and running tests.
Step 5: Review the command logs and test results it returns, then inspect the proposed changes.
Each cloud task runs in a separate environment preloaded with your repository, where the agent can read and edit files and run tests. The CLI runs locally in your terminal.
Most tasks take between 1 and 30 minutes. Codex returns command logs and test results so you can check what it did.
It initially had no general internet access for security reasons. OpenAI later enabled optional internet access during task execution.
It works through a Visual Studio Code extension and integrates with JetBrains IDEs and Xcode. GitHub also offers it in public preview for assigning to issues and pull requests.
Official Codex repository skill for running a final code review on a pull request.
OpenAI DotsDots are always-on agents in ChatGPT that take on ongoing responsibilities and keep making progress between conversations. You choose which apps they use and when they need your approval.
ChatGPT WorkChatGPT Work is an agent in ChatGPT that takes action across your apps and files and turns a goal into finished work such as sheets, slides, docs and web apps. It can stay with a project for hours.
ChatGPT SpaceChatGPT Space is a shared place in ChatGPT for pages, files, spreadsheets and slides that you build with ChatGPT and your teammates. It replaces Library and is available on Pro, Business and Enterprise plans.
Browser UseBrowser Use is an AI-focused agent for automation workflows, giving users practical model-powered capabilities through a dedicated product or developer interface.