Emergent launches Wingman — an autonomous assistant for developers

Emergent, a vibe-coding application creation company, has unveiled Wingman, an autonomous agent capable of managing applications for daily tasks. The company states in its press release, "The best technology should be accessible to everyone," highlighting the challenges faced by users without a technical background in creating software applications.

Wingman enables users to deploy a team of agents working on their behalf. "Now, anyone can have an always-on team working in the background, not just those who know how to build one," said Mukund Jha, co-founder and CEO of Emergent. Wingman distinguishes itself from similar platforms by delineating tasks that can be performed without human intervention from those requiring human approval.

Tasks such as modifying or deleting data, or sending messages to groups, are put on hold until the AI receives the go-ahead from its operator. The company defines these separations as “trust boundaries.” The platform can interact with common applications like WhatsApp, Telegram, and iMessage, and can schedule tasks or trigger them based on preset events.

A window of persistence (short-term context) means that users do not have to repeat contextual instructions for similar tasks. Connections to familiar platforms such as email, calendaring, CRMs, and GitHub are available out of the box, with additional integrations accessible from the company’s hub. In line with the platform’s easy-to-use ethos, connections between Wingman and other applications are established without the need for coding elements like API calls and key exchanges.

Responses from Wingman can be adjusted in tone, allowing it to feel like “a trusted operator rather than just another management tool,” as stated in Emergent’s press release. Wingman operates on a choice of LLMs, including the latest models from ChatGPT and Anthropic, or users can opt for Emergent’s own AI instance to save costs. Sign-up is quick and straightforward, and users can select the development of full-stack or mobile apps or have the AI design web pages.

Plans are available for $20 or $200 per month if paid monthly, with introductory discounts for those wanting to experiment with having an LLM act on their behalf through the applications they use daily. Apps are built using modern, web-native technologies for a professional front end to the resulting code. “Most people aren’t failing at productivity. They’re buried under the smaller tasks that never stop coming,” Jha noted.

The promise of Emergent’s Wingman and similar offerings lies in empowering the true ‘citizen developer,’ where all that is required from the business founder is the ability to articulate their software needs in their native language. The large language model works to interpret those needs using a body of data gathered by scraping the internet for existing code. This is then reproduced, partially randomized, and subtly altered to align closely with the user’s goals.

While tools like OpenClaw and Wingman may be suitable at this stage for hobbyists with specific problems to solve, releasing software created in this manner for broader consumption raises some debatable assumptions about its inherent security and veracity—elements of the final creation that, while readable, will be impenetrable for the intended market of the platforms.