Key Takeaways
- Perplexity has reportedly built an internal coding tool called Teammate, and its engineers have been using it since May.
- The tool is described as aimed at long-horizon engineering work, including owning projects and monitoring services.
- A public launch has not been confirmed, and it is still unclear whether or when the product will ship.
What happened
Perplexity, best known for its AI search engine, appears to be exploring a new product area: AI coding tools. According to a report from Business Insider cited by The Next Web, the company has developed an internal tool codenamed Teammate and may eventually release it publicly.
The report says Perplexity’s engineers have been using Teammate since May. An internal announcement viewed by Business Insider describes the tool as being designed for “long-horizon engineering work,” including owning projects, investigating issues, and monitoring services.
Screenshots reportedly show engineers using the system on real tasks, including bug hunting inside internal systems. The report also says Teammate is model-agnostic, meaning it does not rely on a single chatbot or model family.
Perplexity, which reached a $20 billion valuation in a funding round last year, declined to comment on the report. It is still not clear whether Teammate will ship publicly or on what timeline.
Why it matters
If Perplexity does move forward, Teammate would put the company into a fast-growing and increasingly crowded market. AI coding tools have become one of the clearest commercial use cases for generative AI, drawing heavily funded players such as Cursor, Anthropic, and OpenAI.
That makes Perplexity’s reported interest notable for two reasons. First, it suggests the company is looking beyond search and into workflow software where AI can directly assist engineering teams. Second, the model-agnostic approach may distinguish Teammate from tools built tightly around a single vendor’s own models.
The broader signal is that AI companies are continuing to push from chat and search toward more operationally useful products. A tool that can help manage software projects, investigate bugs, and monitor services is positioned as more than a code autocomplete system; it is closer to an engineering assistant.
The report also highlights how much of the AI coding market now revolves around tools that can already generate real revenue. That makes the category attractive, but it also raises expectations around reliability, quality checks, and practical usefulness.
What to watch
The main question is whether Perplexity chooses to launch Teammate publicly, and if so, how it positions the product against existing coding agents and developer tools.
It will also be worth watching whether Perplexity continues to emphasize a model-agnostic setup, since that would set it apart from tools more closely tied to their creators’ own models.
Finally, the company’s internal messaging is a clue to its direction. The report says CTO Denis Yarats told engineers to “stop looking at code” and rely on AI by year’s end or sooner, while also arguing that generated code can avoid becoming “slop” if it passes quality checks. Whether that philosophy shows up in a product release remains to be seen.
