Parallel vs Valyu

Valyu is best suited for AI knowledge work that combines specialist datasets, web search, covered licensed publications and a firm's knowledge for research and deepresearch. DeepResearch turns those inputs into cited reports and structured outputs. Parallel is known for web search and research APIs built for AI agents.

TL;DR

Choose Valyu when the assignment needs named financial, academic, biomedical and scientific datasets, internal knowledge base (in any format) and a generated research document. Parallel focuses on web search and research for agents. Valyu brings the source choices, document inputs and saved research brief into one task.

Parallel vs Valyu at a glance

DeepResearch benchmark results

In Valyu's published DRACO evaluation, Valyu DeepResearch (Heavy) scored 72.7%, compared with 51.2% for Parallel Ultra8x — a 21.5-percentage-point lead. DRACO grades 100 long-form research tasks across 10 professional domains against expert-written rubrics. The results are reported in Introducing DeepResearch.

DRACO research scores and cost per 1,000 requests, from Valyu's published launch evaluation.

For the deal-review example below, useful research must compare the forecast with reported figures, separate seller claims from evidence and leave open diligence questions visible. The benchmark tests how well a finished report meets an expert rubric, making it a useful complement to the document inputs and report formats described here. View the methodology, raw outputs and evaluations.

Financial Research With Deal Documents

An acquisition team has the seller's presentation, a forecast workbook and its own investment criteria. Before the committee meeting, it needs to check which growth and margin assumptions have supporting evidence and which depend on the seller's projections.

Valyu's financial sources include SEC filings, financial statements and earnings data. Attach the deal files, select the external records and request the memo structure the committee uses:

Review the attached seller presentation and forecast against our investment criteria. Compare the growth and margin assumptions with supporting company figures and relevant public peers, explain adjustments, and return a cited investment memo separating reported facts, seller claims, our assumptions and open diligence questions.

Analysts can inspect the cited comparisons, challenge unsupported assumptions with the seller and revise the model before the meeting. The memo gives the committee the findings and unresolved questions in one review document.

Search the literature behind the method

Covered Wiley finance papers and book chapters provide passages explaining valuation and modelling methods. An analyst can check the method behind an assumption as well as the figures used in the model.

Source access depends on the dataset and account permissions. Confirm that the paper or chapter required for the assignment is included in the collection.

Make the research brief reusable

Valyu DeepResearch lets teams define the question, select search sources and specify the report structure. It conducts the investigation and returns a cited report or structured data. PDF output is supported; generating Word documents or Excel workbooks requires code execution.

Save a useful brief through the Workflows API and reuse it with new inputs. Workflows is in Beta. A pinned version preserves the instructions, while live sources and findings can still change.

Bring the firm's own knowledge into the research

Attach supported PDFs, Word documents or Excel workbooks to an individual DeepResearch task. Tell the agent which assumptions to check and how the documents relate to the external evidence.

For continuing archives, Valyu offers enterprise private indexes and Forward Deployed Engineers who configure retrieval around the team's work. Past memos, lab reports and portfolio reporting can contribute to the same investigation.

Which should you choose?

Parallel is a fit for web search and research APIs for agents. Choose Valyu when the firm's knowledge work requires specialist datasets, covered licensed literature, original deal files and cited reports or reusable review templates.

Use the next deal packet and the committee's actual questions. Attach the documents, select the required sources and inspect the cited figures and outstanding diligence questions before review.

Try Valyu or talk to the team about deal documents, Wiley access and a private index. For an existing agent, use the MCP setup guide.