Grass Launched API to Scraping Web Data for AI
The new tool routes web requests through residential proxies to feed live data to AI agents.
Updated on Oct. 8, 2026 in Artificial Intelligence
Live Poll
Do you trust AI information sourced from real-time web data over pre-trained data sets?
Grass has released its Contents API, a service designed to convert public websites into model-ready formats like Markdown and HTML. The network utilizes a distributed pool of residential internet connections to bypass restrictions that often block data-center traffic.
Why it matters
As AI agents increasingly require real-time information, conventional data infrastructure has struggled to navigate websites that restrict standard scraping techniques. This shift toward residential-routed traffic aims to provide the reliable, up-to-date data necessary for active inference models.
The Contents API achieved a 98% retrieval success rate in hard-domain tests, a 27-point performance lead over the 71% success rate observed in the closest competitor.
The players
Grass
A decentralized data network built on the Solana blockchain that leverages residential proxy bandwidth to scrape web information.
Multicoin Capital
A thesis-driven investment firm focused on crypto-native companies and decentralized infrastructure projects.
Andrej Radonjic
The CEO of Grass who is overseeing the company pivot toward live inference-time data services.
The details
The platform functions by routing HTTP data requests through residential IP addresses instead of traditional data centers. This method allows the API to circumvent anti-scraping filters that typically flag and block enterprise traffic. Once a request is fulfilled, the system converts the returned raw data into a structured output—HTML, Markdown, or plain text—ready for immediate ingestion by large language models.
Timeline
2025: Grass reported $17 million in annual revenue.
First half of 2026: Grass generated $17 million in revenue.
September 28, 2026: CEO Andrej Radonjic announced a company-wide roadmap shift.
September 29, 2026: Multicoin Capital expanded its investment position in Grass.
October 8, 2026: Grass officially launched the Contents API.
The Tech Race
The launch of the Contents API signals an aggressive move to secure a lead in the specialized market for AI training and inference data. By moving beyond traditional scraping, the firm is positioning itself to capture a larger share of the projected $65 million to $75 million in full-year 2026 training-data revenue.
Developers and AI researchers can now integrate the Contents API to automate the conversion of complex web pages into machine-readable text. The service aims to replace manual data cleaning workflows, though pricing and enterprise-tier availability details remain unannounced.
The takeaway
The pivot to live inference-time data indicates that real-time web access is becoming as critical to AI agents as initial training sets. Watch for the company's final 2026 revenue figures to see if they meet the current $65 million to $75 million projection.
Further reading
For more on how infrastructure is evolving to meet the needs of automated systems, explore our Artificial Intelligence section.
Source note: This article includes information reported by Crypto Briefing.
Live Poll
Do you trust AI information sourced from real-time web data over pre-trained data sets?
