AGENTS THAT ANSWER — AND RESEARCH

From instant, cited answers to deep, multi-step investigations with real deliverables.

Not every question needs the same depth

Sometimes your agent needs a fast, grounded answer. Sometimes it needs to investigate, synthesize, and produce a report you can actually use. Valyu Agents handle both — using the same search infrastructure, the same source control, and the same commitment to traceability.

Try Valyu

Answer API

Chat ViewAPI View

Meta-analyses of immunotherapy efficacy in lung cancer

ResearchClearRun⌘⏎

Answer

Network-meta-analyses across the NSCLC spectrum consistently show that immune-checkpoint inhibition improves survival outcomes, especially when added to chemotherapy. In the adjuvant setting for resected early-stage disease, a recent NMA of 19 RCTs (9,438 patients) found that ICI-based regimens lengthened disease-free survival (median DFS ≈ 53 months) compared with chemotherapy or placebo, and both EGFR-TKIs and ICIs trended toward overall-survival benefit, albeit with a higher risk of severe adverse events (Nature). For advanced NSCLC, multiple NMAs of first-line trials report pooled hazard ratios of 0.62–0.74 for progression-free survival and 0.71–0.79 for overall survival when ICIs (pembrolizumab, atezolizumab, nivolumab, etc.) are combined with platinum-based chemotherapy, with an objective-response-rate increase of ≈1.5-fold and a modest rise in grade ≥ 3 toxicities (Ascopubs ; doiDoi). Direct comparisons between ICI monotherapy and chemo-ICI in high-PD-L1 disease show superior response (OR ≈ 1.7) and progression-free survival (HR ≈ 0.59) for the combination, while overall-survival differences are modest (Nature).

Deep Research API

Chat ViewAPI View

(beta)

Analyze the legal framework governing commercial space activities. Examine the Outer Space Treaty and subsequent agreements, catalog national space laws (US Commercial Space Launch Act, Luxembourg space mining law, UAE space law), document liability frameworks for space debris and collisions, analyze property rights for space resources (asteroid mining, lunar resources), evaluate licensing requirements for launches

ExamplesClearRun⌘⏎

Deep Research Response

View ReasoningPreviewJSON

CompletedStandard

Used 6 steps to complete the task

Deliverables (1)

a fun visual report ...

PPTX

Output

The Legal Framework Governing Commercial Space Activities: A Comprehensive Analysis of Treaties, National Laws, and Emerging Challenges

Executive Summary

The international legal framework for commercial space activities operates through a dual structure of aging treaties and divergent national laws, creating fundamental tensions between innovation and liability, property rights and non-

Sources Read

887

Words Read

3741406

Hours Saved

280.7

Deep Research API

Chat ViewAPI View

(beta)

ExamplesClearRun⌘⏎

Deep Research Response

View ReasoningPreviewJSON

CompletedStandard

Used 6 steps to complete the task

Deliverables (1)

a fun visual report ...

PPTX

Output

The Legal Framework Governing Commercial Space Activities: A Comprehensive Analysis of Treaties, National Laws, and Emerging Challenges

Executive Summary

Sources Read

887

Words Read

3741406

Hours Saved

280.7

How Deep Research Works

1. Define the Task

Natural language input. Choose depth, sources, and outputs.

2. Investigation

Planning, searching, extraction, and synthesis — fully automated.

3. Receive Deliverables

Download files or consume structured outputs with citations.

How Deep Research Works

1. Define the Task

Natural language input. Choose depth, sources, and outputs.

2. Investigation

Planning, searching, extraction, and synthesis — fully automated.

3. Receive Deliverables

Download files or consume structured outputs with citations.

Research that produces work-ready outputs

Knowledge work doesn’t stop at answers. It ends with reports, analyses, and documentation. Deep Research plans and executes multi-step investigations, synthesizes findings across sources, and generates deliverables you can use immediately — from PDFs to spreadsheets to structured data.

Source Control

Fine-grained control across web, academic, financial, and medical sources.

File Inputs

Upload PDFs, images, and documents to extend the research context.

Custom Schemas

Define exactly how results should be structured.

Full Traceability

Every insight links back to its source.

Try Deep Research

DeepResearch API Playground

Enter your research question...

Workflows

ClearRun⌘ ⏎

New Chat

History

QUICK SETTINGSDocsD

Mode

Fast

~5 min

Standard

~10-20 min

Heavy

~30-45 min

Max

Up to 2hrs

Response Format

Markdown

JSON

TOON

Deliverables

ExcelCSVSlidesWordPDF

CompletedStandard

View ReasoningPreviewJSON

Used 5 steps to complete the task

OUTPUT

mRNA Therapeutic Platforms Beyond COVID-19: Comprehensive Pipeline Analysis and Commercial Outlook

Executive Summary

The global mRNA therapeutic landscape has expanded dramatically beyond COVID-19 vaccines, with 45+ programs in development across oncology, infectious diseases, rare genetic disorders, and cardiovascular conditions Medrxiv. The therapeutic promise extends far beyond pandemic applications: personalized cancer vaccines demonstrate 49% reduction in recurrence or death in melanoma adjuvant settings Cancernetwork, RSV mRNA vaccines achieved FDA approval in 2024-2025 with 83.7% protection efficacy Cancernetwork, and Sanofi's H5N1 mRNA influenza vaccines are dosing patients as low as 0.30 mg/kg in early-phase trials Reddit Wikipedia Cancernetwork. Critical program milestones define 2026-2027 as an inflection point: Moderna's mRNA-4157/V940 (Intsmeran autogene) expects Phase 3 interim readout mid-2026 with formal data September 2026, positioning it for regulatory approval by late 2026 or early 2027 Reddit; BioNTech and Roche's BNT116 (NSCLC off-the-shelf vaccine) Nih

SOURCES (57)View All

The Promise of mRNA Cancer Vaccines: Potential Lives Saved and Economic Value in the U.S.

Sources Read

111

Words Read

236867

Hours Saved

20.7

When to Use Deep Research

01 / 04

Investment and market research

Analyze markets, companies, and trends across financial filings, news, and proprietary data — with outputs ready for decision-making.

02 / 04

Regulatory and compliance analysis

Track regulatory changes, interpret requirements, and assess impact across jurisdictions using authoritative sources.

03 / 04

Scientific literature reviews

Synthesize findings from academic papers, clinical trials, and journals into structured, citable summaries.

04 / 04

Competitive intelligence and strategy work

Map competitors, positioning, and market dynamics to inform strategic planning and execution.

Deep Research API

Examples

ClearRun⌘⏎

DELIVERABLES

Outputs built for real workflows

Try Deep Research

npm install valyu-js

import { Valyu } from 'valyu-js';
 
const valyu = new Valyu("*********************************");
 
// Create deep research task
const response = await valyu.deepresearch.create({
  input: "Analyze the competitive landscape of AI search APIs",
  mode: "standard",
  outputFormats: ["markdown", "pdf"]
});
 
if (!response.success) {
  console.error("Error:", response.error);
} else {
  console.log("Task created:", response.deepresearch_id);
 
  // Wait for completion with progress updates
  const result = await valyu.deepresearch.wait(
    response.deepresearch_id!,
    {
      onProgress: (status) => {
        if (status.progress) {
          console.log(`Progress: ${status.progress.current_step}/${status.progress.total_steps}`);
        }
      }
    }
  );
 
  console.log(result.output);
}

How Answer works

1. Ask

Send a natural language question. Add source filters if needed.

2. Retrieve + Reason

Valyu searches authoritative sources and synthesizes the findings.

3. Answer

Get clear responses with citations and underlying search results.

How Answer works

1. Ask

Send a natural language question. Add source filters if needed.

2. Retrieve + Reason

Valyu searches authoritative sources and synthesizes the findings.

3. Answer

Get clear responses with citations and underlying search results.

Search + reasoning in a single call

When you don’t need raw search results, you don’t have to manage them. Answer mode runs Valyu search, evaluates the sources, and returns a grounded response with inline citations — fast enough for real-time agent workflows.

Source Control

Restrict answers to specific datasets or domains.

Structured Responses

Define a JSON schema. Get answers in exactly the format your app needs.

Inline Citations

Every claim traces back to a source.

Fast by Default

Most answers return in under 30 seconds.

Try Answer

Answer API

ChatAPI

Meta-analyses of immunotherapy efficacy in lung cancer

ClearRun⌘⏎

Included Datasets

Clear

Search datasets...

All CategoriesResearch & AcademicHealthcare & MedicalPatents & IPFinancial MarketsCompany DataEconomic IndicatorsPrediction MarketsTransportation

Healthcare & Medical

Clinical Trials

DailyMed

ChEMBL

PubChem

DrugBank

Open Targets

API Response

Answer

Recent meta‑analyses consistently show that immune‑checkpoint inhibition improves survival across the lung‑cancer spectrum. In the adjuvant setting for early‑stage non‑small‑cell lung cancer (NSCLC), a network meta‑analysis of 19 randomized trials reported that immune‑checkpoint inhibitor (ICI) therapy lengthened disease‑free survival (DFS) to a median of 53.47 months versus 39.49 months with chemotherapy or placebo (Nature).

In advanced disease, pooled data from randomized trials demonstrate a robust overall‑survival (OS) advantage for immunotherapy over chemotherapy. Real‑world French data showed median OS of 16.4 months with first‑line ICIs versus 11.6 months with chemotherapy (Nature). Trial‑level meta‑analyses of PD‑1/PD‑L1 plus CTLA‑4 combinations reported a 69 % reduction in the risk of progression (HR 0.69, 95 % CI 0.56–0.85) (Medrxiv). Dual‑immunotherapy plus chemotherapy further increased median OS to ≈15.8 months (Cureus).

Biomarker‑focused meta‑analyses indicate that high tumor‑mutational burden (TMB) predicts greater immunotherapy benefit (Nih). Together, these systematic reviews confirm that immunotherapy provides significant survival benefits.

Show Code SnippetsM

When to Use Answer Mode

01 / 04

Research assistants and copilots

Answer complex questions with grounded responses and traceable sources.

02 / 04

Customer support and internal Q&A

Deliver consistent, accurate answers grounded in documentation and knowledge bases.

03 / 04

Fast fact-checking with citations

Verify claims instantly with authoritative sources and inline references.

04 / 04

Real-time agent responses

Power live agents with fast, reliable answers backed by search.

Answer API

Examples

ClearRun⌘⏎

SoTA across several benchmarks

DRACO

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.

Valyu DeepResearch achieved 72.7% accuracy on the DRACO long-form research benchmark, outperforming Perplexity Deep Research (70.5%), Claude Opus 4.6 (59.8%), Gemini Deep Research (59.0%), and OpenAI Deep Research o3 (52.1%), while operating at less than half the cost of the next-best system.

DRACO

Accuracy (%)

VALYU DEEPRESEARCH

72.7% / $2,500CPM

PERPLEXITY

CLAUDE 4.6

GEMINI

YOU.COM

OPENAI O3

PARALLEL

CLAUDE 4.5

OPENAI O4-MINI

TAVILY

EXA

Cost (CPM)

Cost vs accuracy across deep research APIs

$10

$100

$1,000

$10,000

Valyu DeepResearchHeavy

72.7% / $2,500 CPM

Perplexity DR (Opus 4.6)

70.5% / ~$5,500 CPM

Claude Opus 4.6 (web+code)

59.8% / ~$5,500 CPM

Gemini Deep Research

59.0% / ~$4,500 CPM

You.com Researchexhaustive

Claude Opus 4.5 (web+code)

Accuracy (%)

Cost per thousand requests (CPM)

Integrate in seconds

from valyu import Valyu
 
valyu = Valyu(api_key="*********************************")
 
response = valyu.answer(
    "your question here"
)
 
# Access the answer
print("Answer:", response.contents)
print("Sources used:")
for source in response.search_results:
    print(f"- {source.title}: {source.url}")

Build agents withaccess to knowledge

Get API Key