Pydantic AI Integration

Scavio ships as a search tool for Pydantic AI via the pydantic-scavio package. Install it, add scavio_search_tool() to an Agent, and the model gets live Google results as clean, typed JSON — a cost-effective Tavily and SerpAPI alternative at 1 credit per search.

Introduction

The pydantic-scavio package wraps Scavio's Google Search (v2) endpoint as a Pydantic AI Tool. Before you start you need:

1. Pydantic AI: pydantic-ai or pydantic-ai-slim (Python 3.10+).
2. A Scavio API key from dashboard.scavio.dev (50 free signup credits, no credit card).

Step-by-Step Integration Guide

Step 1: Install

pip install pydantic-scavio

Step 2: Set your API key

export SCAVIO_API_KEY=sk_live_your_key

Step 3: Basic usage

import os

from pydantic_ai import Agent
from pydantic_scavio import scavio_search_tool

agent = Agent(
    "openai:gpt-5.5",
    tools=[scavio_search_tool(os.environ["SCAVIO_API_KEY"])],
    system_prompt="Search the web when you need current information.",
)

result = agent.run_sync("What changed in the latest Pydantic AI release?")
print(result.output)

Fixing parameters for the agent

Parameters passed to the factory are fixed for every search and hidden from the LLM's tool schema. Parameters left unset stay available for the model to set per call.

# Always search from the US in English; the LLM only controls the query.
tool = scavio_search_tool(
    os.environ["SCAVIO_API_KEY"],
    country_code="us",
    language="en",
)

Sharing a client

Pass an existing AsyncScavioClient to share connection pooling and rate limiting across tools:

from scavio import AsyncScavioClient
from pydantic_scavio import scavio_search_tool

client = AsyncScavioClient(api_key=os.environ["SCAVIO_API_KEY"])
tool = scavio_search_tool(client=client)

Result shape

Use every endpoint via MCP

For the full Scavio API (YouTube, Amazon, Walmart, Reddit, TikTok, Instagram, and all Google surfaces) with no extra install, point Pydantic AI at the hosted MCP server:

from pydantic_ai import Agent
from pydantic_ai.mcp import MCPServerStreamableHTTP

server = MCPServerStreamableHTTP(
    url="https://mcp.scavio.dev/mcp",
    headers={"x-api-key": "sk_live_..."},
)
agent = Agent("openai:gpt-5.5", toolsets=[server])

Benefits of Scavio + Pydantic AI

• Typed end to end: validated results that match Pydantic AI's type-safe philosophy.
• Schema control: developer-fixed parameters never leak into the LLM tool schema.
• Cost-effective: 1 credit per search, full results.
• One key, every platform: the same account covers 46 MCP tools across 7 platforms.

Next Steps

• Pydantic AI documentation
• pydantic-scavio on GitHub
• Scavio API quickstart