Startup Parallel Web Systems Inc. today launched a cloud platform that artificial intelligence applications can use to conduct online research.

The two-year-old company is led by former Twitter Inc. Chief Executive Parag Agrawal. Parallel is backed by Khosla Ventures, First Round Capital, Index Ventures and several other investors. The Information reported last January that the company had raised $30 million in funding.

Parallel’s platform enables AI models to incorporate data from the public web into prompt responses. According to the company, its software already completes millions of research tasks per day for early adopters. Those customers include “some of the fastest growing AI companies,” Parallel staffers detailed in a blog post published today.

The company says that its platform lends itself to a wide range of use cases. An AI programming assistant, for example, could harness the software to pull code from GitHub repositories. Retailers can use it to collect information about a rival’s product catalog.

The platform provides access to eight AI research engines with varying capabilities. The fastest, most cost-efficient engine can process requests in under one minute. Parallel’s most capable engine, Ultra8x, retrieves more detailed information and takes up to 30 minutes to complete a single research task.

The platform’s prompt responses include confidence scores that help customers assess the quality of retrieved data. Additionally, Parallel provides citations that make it easier to verify the data.

Customers can optionally have the platform organize the information that it gathers in a specific format. A company could, for example, ask Parallel to organize product reviews from competitor websites in a three-column spreadsheet. It’s also possible to customize other details, including the amount of computing capacity that the platform uses to complete each research task.

Parallel compared its platform with OpenAI’s GPT-5 ahead of today’s launch. The evaluation used BrowseComp and DeepResearch Bench, two benchmarks that test LLMs’ ability to conduct online research. Parallel’s top-end Ultra8x research engine outperformed GPT-5 by more than 10% across both benchmarks. 

AI models can access the company’s platform through three application programming interfaces. There’s a general-purpose Task API, as well as a Search API that is specifically optimized to power AI agents. Parallel’s third API is designed for chatbots and other applications that prioritize low latency.

In today’s launch blog post, the company indicated that it’s working to expand the range of use cases that its platform supports. The company hopes to facilitate the creation of AI agents that can “complete the work of entire teams in hours.” Additionally, it plans to give agents the ability to automatically perform actions in response to webpage changes.

Image: Parallel