The answer existed. It was buried in 15 hours of searching.

The company’s scientists regularly need to know if existing research already answers a question before investing time and resources into new studies. Does mycorrhiza help protect raspberry plants from ground freezing? Is there documented evidence on a particular soil treatment? What does the literature say about a specific growing condition?

The process to answer those questions was entirely manual. A scientist would open Google Scholar, build a search query, scan through results, read abstracts, filter irrelevant papers, cross-reference sources, and eventually synthesize what they found into something usable. For a single question, that process routinely took 10 to 15 hours, and was limited to a single source. Relevant research sitting in other scientific libraries simply went unseen.

The cost wasn’t just time. It was the risk of making decisions without the full picture, of duplicating research that already existed, or of missing a critical finding because it lived outside of Google Scholar.

Mirego built a research agent on Forra that handles the entire front-end of the scientific research process, from query to structured PDF output. The scientist asks a question in plain language. The agent takes it from there.

Rather than querying a single database, the agent searches across four scientific libraries simultaneously, casting a net wide enough to surface research that would have been missed entirely in a manual process. It doesn’t just retrieve results: it filters for relevance, reads and understands the content, and synthesizes the findings into a structured research report delivered as a ready-to-use PDF.

The scientist’s role shifts from searcher to reviewer. Instead of spending a day building a literature review from scratch, they receive a complete, sourced synthesis in minutes and can focus their expertise on interpreting the findings and deciding what to do next.