Pre-built Agent

AI Search Agent

AGENT ID: 8eb884a7-b68a-4be0-932c-6904548798c5

It is an AI search agent that can search the internet in real time using an LLM to answer user queries.

Deploy / Run Agent

API & SDK Integration

Load, run, and interact with this agent programmatically using standard HTTP curl requests or the SVAHNAR SDK.

Secure Environment

Runs in isolated, zero-retention execution nodes with encrypted data transport.

Models Used (1)

gpt-4o-mini

Tools Attached (1)

Tavily

MCP Connections (0)

No Model Context Protocol servers connected.

Specifications

AGENT TYPE

Single Agent

SUB-AGENTS

1 agent

SUPERVISOR

RUNS

1,200+ runs

Configuration (YAML)

create_vertical_agent_network:
  agent-1:
    tools:
      tool_assigned:
      - name: Tavily
        config:
          API_key: ${TAVILY_API_KEY}
    LLM_config:
      params:
        model: gpt-4o-mini
        max_tokens: 1000
        temperature: 0.5
        request_timeout: 600
    agent_name: Agent_1
    incoming_edge:
    - Start
    outgoing_edge: []
    agent_function:
    - You are a search agent whose task is to search for user query on web or internet
      and grab the searched result. It should be relevant, proper and appropriate

Developer Quick Start

1. Install the SDK

pip install svahnar

2. Configure your API key

export SVAHNAR_API_KEY="your_api_key_here"

3. Save the Configuration YAML

Copy the YAML configuration box above and save it locally as agent.yaml.

4. Deploy the Agent (Python)

from svahnar import Svahnar
from pathlib import Path

client = Svahnar()

# Deploy the agent configuration
agent = client.agents.create(
    name="AI Search Agent",
    description="It is an AI search agent that can search the internet in real time using an LLM to answer user queries.",
    deploy_to="Organization",
    yaml_content=Path("agent.yaml")
)
print(f"Created Agent ID: {agent.id}")

5. Run the Agent (Python)

from svahnar import Svahnar

client = Svahnar()

# Run the agent using its ID
response = client.agents.run(
    agent_id="YOUR_DEPLOYED_AGENT_ID",  # Replace with your agent ID from Step 4
    message="Your input message here"
)
print(response)

For advanced features, Human-in-the-Loop configurations, and guides, visit the Developer Documentation.