How to Use the Sensible MCP in AutoGen

Q: How can I set up two AutoGen agents to debate the best Sensible extraction method?

Create two AssistantAgent instances, each with a different system message. One's prompt could prioritize speed (favoring extractsync), the other accuracy (favoring extractsyncwithconfig). Then, have them discuss which tool to use for a given document.

Q: Can one AutoGen agent create a Sensible configuration and another one publish it?

Absolutely. You can give the createconfiguration tool to a 'developer' agent and the publishconfiguration tool to a 'releasemanager' agent. The release manager can then refuse to act until it gets a success message from the developer.

Q: What's a good agent pattern for handling asynchronous jobs from Sensible with AutoGen?

Have one agent call extractfromurl and pass the job ID to a second 'polling' agent. The polling agent can then periodically call getdocument until the status is 'completed', at which point it passes the final result back to the group chat.

Q: Is it possible to have an AutoGen agent that only has permission for read-only Sensible tools?

Yes. When you generate the tool list using mcpservertools, you can filter it before passing it to your AssistantAgent. Simply provide the agent with a list containing only read-only tools like listextractions or getdocumenttype.

Assemble a team of AutoGen agents to collaborate and debate on complex document parsing tasks.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers — Included with Plan

Connect Sensible MCP to AutoGen

Create your Vinkius account to connect Sensible to AutoGen — we handle the hosting, security, and runtime updates so you don't have to. No server setup required.

GDPR Included with Plan

Setup Sensible with AutoGen

Ask AI about this MCP

ChatGPT

Claude

Perplexity Vinkius runs on Gemini

Gemini

Grok

Multi-Agent Configuration Management

Let your agents build and validate parsing configurations together. One agent, a "Schema Specialist," can propose a new document structure using `create_document_type` and `create_configuration`. A second "QA Bot" agent can then test it against a known-good file using `create_golden` and `extract_sync_with_config`. Only when the QA Bot confirms the output is correct will a third "Release Manager" agent be allowed to call `publish_configuration`. This multi-agent workflow catches errors before they hit production. It turns a risky manual process into a supervised, automated debate.

Debate Extraction Strategies with a Sensible MCP Server

Don't just pick one extraction method—let your agents argue for the best one. A "Speed-Focused" agent might advocate for using `extract_sync` for immediate results. A "Reliability-Focused" agent could counter that `extract_from_url_with_config` is more robust for known document types. They can even pull data to support their arguments, for example, by checking `get_extraction_statistics` to see which method has a better success rate. Your user gets a final answer that has been challenged and defended by multiple specialized agents.

Collaborative Document Review

Use an agent conversation to manage document analysis. A "Processing" agent can run an extraction with `extract_from_url` and pass the results to a "Reviewer" agent. If the Reviewer spots a problem, it can flag it and assign the job back to the Processor. This loop can even include a human. The Reviewer agent can use `get_auth_tokens` to generate a temporary login for a person to visually inspect a difficult document. This creates a full feedback loop where AutoGen agents and humans work together through the MCP.

Setup guide

Set up Sensible MCP in AutoGen

Prerequisites

Python 3.10+ installed
autogen-ext[mcp] package
Active Vinkius subscription with a valid endpoint token

1

Install AutoGen with MCP
Run pip install "autogen-ext[mcp]" autogen-agentchat. The MCP extension includes mcp_server_tools for stateless tool access.
2

Fetch tools from the MCP
Call mcp_server_tools(SseServerParams(url=...)) with your Vinkius endpoint. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com.
3

Run your agent
Pass the tools to AssistantAgent and call agent.run(). The agent invokes Sensible tools and returns structured results.

agent.py

from autogen_ext.tools.mcp import SseServerParams, mcp_server_tools
from autogen_agentchat.agents import AssistantAgent
from autogen_ext.models.openai import OpenAIChatCompletionClient

server_params = SseServerParams(
    url="https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
)

tools = await mcp_server_tools(server_params)

agent = AssistantAgent(
    name="Sensible_assistant",
    model_client=OpenAIChatCompletionClient(model="gpt-4o"),
    tools=tools,
)

result = await agent.run("List recent Sensible data")
print(result.messages[-1].content)

Get your connection token →

Prerequisites

Python 3.10+ installed
autogen-ext[mcp] + autogen-agentchat
Active Vinkius subscription with a valid endpoint token

1

Install dependencies
Same packages as above. McpWorkbench is ideal when your agent needs stateful sessions across multiple tool calls.
2

Use McpWorkbench as context manager
Wrap your agent in async with McpWorkbench(...) to maintain shared state and resources. The workbench manages the full MCP session lifecycle.
3

Run with workbench
Pass workbench=workbench to your agent. State is preserved across multiple tool calls within the same session.

agent.py

from autogen_ext.tools.mcp import McpWorkbench, SseServerParams
from autogen_agentchat.agents import AssistantAgent
from autogen_ext.models.openai import OpenAIChatCompletionClient

server_params = SseServerParams(
    url="https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
)

async with McpWorkbench(server_params) as workbench:
    agent = AssistantAgent(
        name="Sensible_assistant",
        model_client=OpenAIChatCompletionClient(model="gpt-4o"),
        workbench=workbench,
    )

    result = await agent.run("List recent Sensible data")
    print(result.messages[-1].content)

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 Sensible. 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 Sensible 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 Sensible MCP in AutoGen

Create two `AssistantAgent` instances, each with a different system message. One's prompt could prioritize speed (favoring `extract_sync`), the other accuracy (favoring `extract_sync_with_config`). Then, have them discuss which tool to use for a given document.

Absolutely. You can give the `create_configuration` tool to a 'developer' agent and the `publish_configuration` tool to a 'release_manager' agent. The release manager can then refuse to act until it gets a success message from the developer.

Have one agent call `extract_from_url` and pass the job ID to a second 'polling' agent. The polling agent can then periodically call `get_document` until the status is 'completed', at which point it passes the final result back to the group chat.

Yes. When you generate the tool list using `mcp_server_tools`, you can filter it before passing it to your `AssistantAgent`. Simply provide the agent with a list containing only read-only tools like `list_extractions` or `get_document_type`.

Your agents pass document data, like the content of PDFs and images, to the Sensible MCP Server for extraction. You control the conversation. Data is shared between your agents and the server according to the logic you define, and Vinkius does not persist it after extraction. In an AutoGen setup, you are responsible for how agents share and expose the extracted data during their collaboration.

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