How to Use the Harvest MCP in Pydantic AI

Q: What's the easiest way to list all unbilled time for a client in Harvest using Pydantic AI?

First, use the listclients tool to get the ID for the client you want to bill. Then, pass that client ID to the listtimeentries tool and filter for entries where isbilled is false. The typed response makes filtering simple.

Q: Does this integration support creating invoice line items?

The createinvoice tool takes a list of time entries to be billed. The MCP server handles turning those entries into the correct line items on the Harvest invoice draft, so you don't have to construct them yourself.

Get type-safe Harvest automation with Pydantic AI. Ensure every client, invoice, and time entry your agent handles is 100% correct.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Harvest MCP to Pydantic AI

Create your Vinkius account to connect Harvest to Pydantic AI and route execution through our secure gateway. The platform manages server hosting, runtime updates, and security layers. Configuration requires no manual server provisioning.

GDPR Free for Subscribers

Setup Harvest with Pydantic AI

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Build Invoices That Can't Be Wrong

Stop writing defensive code to handle weird API outputs. When your agent uses `create_invoice`, the response is automatically parsed and validated against a Pydantic model. You get a clean, typed object to work with. If the server ever returns something unexpected—a missing field, a string instead of a number—your code raises a `ValidationError` on the spot. This makes it impossible for your agent to silently create or pass along a broken invoice.

Manage Clients with Any LLM, Safely

Pydantic AI is model-agnostic, so you can use these Harvest tools with OpenAI, Anthropic, Gemini, or a local model. The important part is that the tools, like `create_client` and `get_client`, are backed by Pydantic's strict data models. No matter which LLM is driving the agent, you have a guarantee that the client data it's working with is correctly structured. Your agent's logic gets simpler because it can trust the data it receives from this MCP Server.

Enforce Correctness with Pydantic AI

This integration is all about data integrity. Every tool call, from `list_time_entries` to `get_invoice`, must return data that conforms to a predefined schema. There are no optional fields unless explicitly defined. You don't have to guess if a time entry will have a project ID or if an invoice has a due date. Pydantic AI validates everything. If the data isn't perfect, your agent fails loudly instead of proceeding with bad information.

Setup guide

Set up Harvest MCP in Pydantic AI

Prerequisites

Python 3.10+ installed
pydantic-ai-slim[fastmcp] package
Active Vinkius subscription with a valid endpoint token

1

Install Pydantic AI with FastMCP
Run pip install "pydantic-ai-slim[fastmcp]". The FastMCP toolset replaces the deprecated MCPServerHTTP class with full protocol support.
2

Configure the FastMCPToolset
Pass a JSON-style config dict to FastMCPToolset with your Vinkius URL. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com. Supports Streamable HTTP, SSE, and Stdio transports.
3

Create and run your agent
Pass the toolset to Agent(toolsets=[toolset]) and call agent.run(). Swap openai:gpt-4o for any supported model — Anthropic, Google, Mistral, or Groq.

agent.py

from pydantic_ai import Agent
from pydantic_ai.toolsets.fastmcp import FastMCPToolset

toolset = FastMCPToolset({
    "mcpServers": {
        "harvest-mcp": {
            "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
        }
    }
})

agent = Agent(
    "openai:gpt-4o",
    toolsets=[toolset],
    system_prompt="You have access to Harvest tools.",
)

result = await agent.run("List recent Harvest transactions")
print(result.output)

Get your connection token →

Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Harvest. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.

Why Choose Vinkius

Vinkius connects your tools to AI with real-time monitoring and automatic cost savings — all from one dashboard.

Connect Harvest now

Real-time monitoring

Live

visibility into every interaction

Connect your favorite tools to your AI and see exactly what's happening — every request, every response, in real time.

Built-in savings

60%

lower AI costs

Vinkius compresses data between your apps and your AI automatically. Lower bills every month — no configuration required.

Single dashboard

One

place for every integration

Every tool your AI connects to, managed from a single screen. One account, complete control.

Common questions about Harvest MCP in Pydantic AI

Pydantic AI validates every single API response from the Harvest MCP server against a strict schema. If the data doesn't match the expected structure, it throws an error immediately, preventing your agent from acting on corrupted or incomplete information.

Yes. Pydantic AI is designed to work with any LLM. You just need to provide a compatible client, and it can use the Harvest tools like `create_client` and `list_invoices` regardless of the underlying model.

First, use the `list_clients` tool to get the ID for the client you want to bill. Then, pass that client ID to the `list_time_entries` tool and filter for entries where `is_billed` is false. The typed response makes filtering simple.

The `create_invoice` tool takes a list of time entries to be billed. The MCP server handles turning those entries into the correct line items on the Harvest invoice draft, so you don't have to construct them yourself.

The server handles your client details, invoice data, and individual time entry records. Vinkius security ensures this data is processed in a single-use, sandboxed container. No data is ever logged or stored; it only exists in memory long enough to serve the request to your Pydantic AI agent.

Use it with your favorite AI tools

Connect this server to Cursor, Claude, VS Code, and more.

OpenAI Agents SDK sdk-python

Google ADK sdk-python

Pydantic AI sdk-python

Vercel AI SDK sdk-typescript