AI

OpenAI debuts Codex CLI, an open source coding tool for terminals

In an attempt to inject AI into more of the programming process, OpenAi is launched Codex CliA coding “agent” that is designed to be performed locally from terminal software.

Codex Cli, announced on Wednesday next to OpenAI’s newest AI models, O3 and O4-Mini, links the models of OpenAI with local code and computer tasks, says OpenAi. Via Codex CLI, the models of OpenAI can write and edit on a desktop and take certain actions, such as moving files.

Codex Cli seems to be a small step in the direction of OpenAi’s wider agent coding vision. Recently, CFO of the company, Sarah Friar, described What she called the ‘Agentic Software Engineer’ is a set of tools that OpenAI is planning to build a project description for an app and to make it effective and even test quality assurance test.

Codex Cli is not going that far. But the shall Integrate the models of OpenAI, ultimately including O3 and O4-Mini, with the clients that process code and computer assignments, also known as command line interfaces.

It is also open source, says OpenAi.

‘[Codex CLI is] A lightweight, open source coding agent who runs locally in your terminal, “a Spokesperson from OpenAi told WAN via e -mail.” The purpose [is to] Give users a minimal, transparent interface to link models directly to [code and tasks]. “

In a blog post to WAN, OpenAi added: “You can get the benefits of multimodal reasoning of the command line by passing on screenshots or low reliability sketches to the model, combined with access to your code locally [via Codex CLI]. “

See also  Instacart CEO Fidji Simo is joining OpenAI

To stimulate the use of Codex CLI, OpenAI is planning to Dole $ 1 million in API subsidies For eligible software development projects. The company says it will assign $ 25,000 blocks of API credits to the chosen projects.

AI coding tools naturally come with risks with risks. A lot of study to have shown These code-generating models often fail to repair and even introduce security vulnerabilities and bugs. It is best to keep that in mind before you give AI access to sensitive files or projects – less entire systems.

Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button