Perplexity is quietly building an AI coding tool

Perplexity wants a piece of the AI coding boom. The San Francisco search startup has built an internal coding tool, codenamed “Teammate,” and may launch it publicly later, a person familiar with the matter told Business Insider. Its own engineers have used the tool since May.

The company made its name with an AI search engine that takes on Google. A coding product would push it onto very different ground, closer to Cursor, Anthropic and OpenAI. All three already run widely used coding tools.

What “Teammate” does

Teammate aims to run software projects end to end, according to an internal announcement that Business Insider viewed. “It’s built for long-horizon engineering work: owning projects, investigating issues, and monitoring services,” the note reads.

Screenshots show engineers handing it real jobs, such as hunting bugs in internal systems. The tool stays model-agnostic, the person said, so it does not sit on any single chatbot. That marks a notable choice, while rivals like Claude Code lean on their makers’ own models.

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

A Perplexity spokesperson declined to comment, and it remains unclear if or when the product ships. The startup reached a $20bn valuation in a funding round last year, so it has room to experiment.

“Stop looking at code”

The push comes from the top. A few weeks before Teammate went live internally, chief technology officer Denis Yarats told engineers that by year’s end, or sooner, they should “stop looking at code” and simply use AI. It echoes OpenAI’s own internal shift to coding agents.

Yarats also pushed back on the charge that AI writes “slop,” or low-quality code. “Slop is not going to be a thing,” he wrote, as long as the generated code passes quality checks.

Why it matters

The timing raises the stakes. AI coding ranks among the few places where the technology already earns real money, which is why the field has drawn so many well-funded entrants. A model-agnostic tool from a $20bn search company would add a fresh angle to a crowded fight.

Two cautions temper it. This report rests on a single source, and Perplexity has not confirmed a launch. But the direction of travel looks clear: even an AI search company now sees writing your code, not just answering your questions, as territory worth taking.