Grass Launched New Contents API for AI Agents

The network has introduced a tool designed to convert live public web data into formats optimized for machine learning.

Updated on Oct. 8, 2026 in Artificial Intelligence

Grass Launched New Contents API for AI Agents

Live Poll

Do you trust AI information sourced from real-time web data over pre-trained data sets?

Grass has officially launched its Contents API, a platform that transforms public web pages into text, HTML, or Markdown for AI models. The system bypasses data-center blocks by routing requests through a distributed pool of residential internet connections.

Why it matters

The company has shifted its focus to providing live, inference-time data for AI agents. This strategy addresses the difficulty traditional data infrastructures face when attempting to scrape websites that actively block data-center traffic.

The Contents API achieved a 98% retrieval success rate during hard-domain testing. This performance represents a 27-point success gap compared to the 71% retrieval rate recorded for the closest competitor.

The players

Grass

Grass is a technology company that operates a decentralized network on the Solana blockchain to facilitate web data retrieval for artificial intelligence.

Andrej Radonjic

Andrej Radonjic is the Chief Executive Officer of Grass who oversees the company's strategic shift toward providing real-time data for AI agents.

Multicoin Capital

Multicoin Capital is a thesis-driven investment firm that recently increased its financial stake in the development of the Grass network.

The details

By utilizing residential internet connections, the network retrieves information that standard web-scraping tools cannot access. The resulting data is delivered in formats tailored for seamless integration into AI model training and live inference.

September 28, 2026: CEO Andrej Radonjic announced a strategic roadmap shift.
September 29, 2026: Multicoin Capital expanded its investment position in Grass.
October 8, 2026: Grass officially launched the Contents API.
2025: The company reported $17 million in revenue.
First half of 2026: The company reported $17 million in revenue.

The Tech Race

The launch of the Contents API follows the pattern of using residential proxy networks to overcome restrictive data-center traffic blocking in AI training. This advancement highlights the ongoing competition to provide AI agents with high-fidelity, real-time data retrieval capabilities.

Users and developers can expect improved access to real-time, clean data for AI models without the common errors associated with traditional scraping. This capability allows for more reliable performance in AI-driven tools that require constant updates from the public web.

The takeaway

As AI agents become more prevalent, the ability to access and format live web data reliably will remain a critical bottleneck for developers. Businesses shifting their infrastructure to support inference-time data processing are likely to see significant growth in training-data revenue.