The biggest pain point for AI agents right now is that they often get blocked when browsing the web.
Here’s one data point: 9.2% of requests from AI bots & crawlers receive an HTTP 403 Forbidden response, and that percentage continued to rise in August and September (source: Cloudflare Radar).
Anyone who’s used these kinds of bots knows the frustration: you ask an agent to look something up in real time, and it gets blocked before you know it.
Grass’s newly launched Contents API is designed to address this problem: it lets AI agents get the information they need directly from the live internet. It launched on October 8, and in official tests, it achieved a 98% success rate at crawling hard-to-access domains. This is also the rationale behind Multicoin increasing its position—the real-time retrieval market in the inference era could be several orders of magnitude larger than the training data market.
This means Grass’s revenue model is about to change: from “user nodes → collect internet data → sell it to AI companies → train models” to “AI agents make real-time requests → Grass network → find and crawl webpages → clean the data → return it to the agent.”
The thing to watch next is whether Grass can distribute a share of this economic value to token holders.
$GRASS
Here’s one data point: 9.2% of requests from AI bots & crawlers receive an HTTP 403 Forbidden response, and that percentage continued to rise in August and September (source: Cloudflare Radar).
Anyone who’s used these kinds of bots knows the frustration: you ask an agent to look something up in real time, and it gets blocked before you know it.
Grass’s newly launched Contents API is designed to address this problem: it lets AI agents get the information they need directly from the live internet. It launched on October 8, and in official tests, it achieved a 98% success rate at crawling hard-to-access domains. This is also the rationale behind Multicoin increasing its position—the real-time retrieval market in the inference era could be several orders of magnitude larger than the training data market.
This means Grass’s revenue model is about to change: from “user nodes → collect internet data → sell it to AI companies → train models” to “AI agents make real-time requests → Grass network → find and crawl webpages → clean the data → return it to the agent.”
The thing to watch next is whether Grass can distribute a share of this economic value to token holders.
$GRASS