Agents that answer and research.
From instant, cited answers to multi-step investigations that deliver reports, sheets and structured data.
How deep research works.
Research that produces work-ready outputs.
Try Deep Research API →Knowledge work doesn’t stop at answers. It ends with reports, analyses and documentation — every figure traced to its source.
This is a multi-faceted query across manager profiles, competitive dynamics, market trends, and sector metrics.
Initial search strategy: launch parallel searches for sector overview, player comparisons, trends, and credit metrics.
Fetch market scale, manager AUM, direct-lending volume, covenant quality, spreads, and redemption stress signals.
The Private Credit Sector: Competitive Landscape, Market Dynamics, and the Structural Shift in Alternative Finance
Private credit expanded from $250 billion in 2008 to $3.5 trillion by 2025, driven by bank regulation, institutional reallocation, and non-bank lending alternatives. [1][2]
Apollo, Blackstone, Ares, and Blue Owl now dominate the sector, controlling roughly 40% of private credit capital through scale, permanent capital, and origination reach. [2][3]
The mega-managers originated approximately $369 billion in direct lending volume across 3,533 deals in 2025, displacing traditional bank lending in the middle market. [2][3]
Defaults are accelerating, covenant protection has eroded, spreads have compressed, and $20.8 billion in Q1 2026 redemption requests marks the first real stress test. [4]
When to use deep research.
Integrate in seconds.
Create a task, wait for completion, receive deliverables — reports, sheets and structured data.
Read the docs →from valyu import Valyu
valyu = Valyu(api_key="$VALYU_API_KEY")
response = valyu.deepresearch.create(
input="Analyze the competitive landscape of AI search APIs",
mode="standard",
output_formats=["markdown", "pdf"]
)
result = valyu.deepresearch.wait(response.deepresearch_id)
print(result.output)State-of-the-art across research benchmarks.
DRACO is Perplexity's open expert-rubric benchmark of 100 long-form deep research tasks across 10 professional knowledge-work domains, including finance, medicine, academic research, law, and technology. Each output is graded by a per-criterion judge against a domain-expert rubric. We ran every commercially available deep research API end-to-end against the same 100 questions. Every search and research API was tested on its highest publicly-available compute tier. Parallel (Ultra8x), You(.)com Research (exhaustive), Tavily (pro), Exa (deep-reasoning), and Perplexity Deep Research (Opus 4.6). Valyu was run on Heavy mode; we offer a higher Max tier but did not use it here, since Max is more expensive per task than the field.
Cost vs accuracy across deep research APIs
