Perplexity builds Teammate AI coding tool to challenge Cursor, Claude Code and OpenAI

3 Sources

Share

Perplexity, known for its AI search engine, has quietly developed an internal AI coding tool called Teammate. Engineers have used it since May for bug investigation and service monitoring. If launched publicly, the $20bn startup would enter direct competition with Cursor, Anthropic's Claude Code, and OpenAI in the lucrative AI coding market.

Perplexity Enters the AI Coding Wars with Internal Tool

Perplexity has quietly built an AI coding tool codenamed Teammate, marking a significant shift for the company best known as an AI search engine

1

. The San Francisco-based startup, valued at $20bn following its latest funding round, has been testing the AI-powered coding assistant internally since May, with its own engineers using the tool for real software development tasks

1

. While Perplexity has not officially announced plans for a public release, the development signals the company's ambition to compete with Claude Code from Anthropic, OpenAI's coding assistants, and Cursor in one of the few AI segments already generating substantial revenue

2

.

Source: Digit

Source: Digit

End-to-End Software Project Management Capabilities

Teammate distinguishes itself through its focus on long-horizon engineering tasks rather than simple code completion. According to an internal announcement viewed by Business Insider, the tool is "built for long-horizon engineering work: owning projects, investigating issues, and monitoring services"

1

. Screenshots reveal that Perplexity engineers have deployed the tool for practical applications such as bug investigation in internal systems

3

. This approach positions Teammate as a comprehensive solution for AI for coding tasks that extend beyond quick fixes to encompass entire project lifecycles.

Source: Analytics Insight

Source: Analytics Insight

Model-Agnostic Architecture Sets It Apart

A notable technical decision makes Teammate different from competing AI-driven coding solutions. The tool operates as model-agnostic, meaning it does not rely on any single chatbot or AI model

1

. This contrasts sharply with rivals like Claude Code, which lean heavily on their creators' proprietary models. The flexibility could allow Perplexity to leverage the best available models for different coding scenarios, though the company has declined to comment on implementation details

3

.

Leadership Push Toward AI-First Development

The initiative comes directly from Perplexity's Chief Technology Officer Denis Yarats, who has urged engineers to embrace AI tools fully. A few weeks before Teammate launched internally, Yarats told the engineering team that by year's end, or sooner, they should "stop looking at code" and simply use AI

1

. This directive mirrors OpenAI's internal shift toward coding agents and reflects growing confidence in AI capabilities for software development

1

. Yarats also addressed concerns about AI-generated code quality, stating that "slop is not going to be a thing" as long as generated code passes quality checks

3

.

Strategic Timing in a Lucrative Market

The move comes as AI coding tools represent one of the few areas where artificial intelligence already generates real revenue, drawing well-funded competitors into an increasingly crowded field

1

. Developers across the industry use these tools to write code, fix errors, and save time, making the market attractive for companies looking to monetize AI capabilities

2

. If Perplexity launches Teammate publicly, it would mark a major expansion beyond its core AI search business and put the company in direct competition with established players like Anthropic and OpenAI

3

.

Infrastructure Investments Signal Serious Intent

Perplexity's commitment extends beyond software to hardware infrastructure. The company reportedly plans to deploy Nvidia Vera CPUs after testing showed the chips completed real coding workflows approximately 1.5x faster than x86 processors, with concurrent sandboxes starting up to 1.9x faster

3

. These performance gains could provide a competitive edge in handling the demanding computational requirements of AI-powered coding assistants. The direction suggests that even companies focused on AI search now view automated code generation as territory worth claiming, though questions remain about timing and market reception for yet another entry in the AI coding assistant space.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved