Real-world agent examples demonstrating Valyu tools for enterprise workflows.
Investment research agent combining SEC filings, market data, and web search.
python financial_analyst.py "Analyze NVIDIA's competitive position"Tools used: financeSearch, secSearch, webSearch
Best for:
- Equity research
- Investment memos
- Earnings analysis
- Sector comparisons
Academic and technical research agent for literature review and prior art search.
python research_assistant.py "transformer architecture improvements"Tools used: paperSearch, patentSearch, webSearch
Best for:
- Literature reviews
- Prior art search
- Technical research
- State of the art surveys
Comprehensive due diligence agent for M&A, investment, or partnership evaluation.
python due_diligence.py "Stripe"Tools used: secSearch, financeSearch, webSearch, patentSearch
Best for:
- M&A due diligence
- Investment evaluation
- Partnership assessment
- Competitive intelligence
-
Set your API key:
export VALYU_API_KEY=your-api-key -
Install dependencies:
pip install valyu-agentcore[strands]
-
Run any example:
python financial_analyst.py "your query"
These examples show patterns you can adapt:
- Tool Selection - Choose tools relevant to your domain
- System Prompts - Define the agent's persona and output format
- Temperature - Lower for factual, higher for creative tasks
- Result Limits - Adjust
max_num_resultsbased on depth needed
Example customization:
from strands import Agent
from strands.models import BedrockModel
from valyu_agentcore import secSearch, financeSearch
agent = Agent(
model=BedrockModel(
model_id="us.anthropic.claude-sonnet-4-20250514-v1:0",
temperature=0.2, # Factual analysis
),
system_prompt="You are a compliance analyst...",
tools=[
secSearch(max_num_results=20), # More SEC results
financeSearch(max_num_results=10),
],
)These agents can also run through AgentCore Gateway. After deploying Valyu to your gateway, the same queries work with centralized auth and audit logging.
See ../gateway/ for gateway setup.