How to Use the Braintrust MCP in CrewAI

Q: How do I add Braintrust tools to a CrewAI agent?

Pass your Vinkius HTTP endpoint directly into the agent's mcps array parameter during initialization. CrewAI automatically fetches the tool schemas and exposes actions like listdatasets to that specific agent.

Q: How do I filter which Braintrust tools a CrewAI agent can use?

Use the MCPServerHTTP class from crewai.mcp and specify a toolfilter. This lets you restrict a junior agent to only run listprompts while preventing it from modifying datasets.

Deploy collaborative agent crews to run Braintrust evaluations and manage prompts autonomously with CrewAI via MCP.

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

MCP Servers - Free for Subscribers

Connect Braintrust MCP to CrewAI

Create your Vinkius account to connect Braintrust to CrewAI 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 Braintrust with CrewAI

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Run collaborative evaluations with CrewAI

The Braintrust MCP Server lets your agents fetch test cases and write results using the `get_dataset` and `create_experiment` tools. One agent can fetch raw test cases, while a separate analyzer agent runs the test and writes the results back. This multi-agent setup uses shared memory to track progress across the entire run. You get structured, coordinated evaluation pipelines without writing complex orchestration scripts.

Automate prompt optimization via MCP Server tools

This Braintrust MCP Server provides the `list_prompts` and `get_prompt` tools so your agents can optimize system instructions. A researcher agent can pull current templates, write variations, and test them against your evaluation datasets. The crew can then save the best-performing templates by creating new versions. This takes the manual trial-and-error out of optimizing model instructions.

Monitor gateway configurations

This Braintrust MCP Server exposes the `list_env_vars` tool to let your agents audit gateway configurations. A monitor agent can verify that your AI gateway has the correct API keys configured for your target models. This prevents silent failures where an evaluation run crashes halfway through due to expired credentials. Your agents can flag configuration issues before starting heavy test suites.

Setup guide

Set up Braintrust MCP in CrewAI

Prerequisites

Python 3.10+ installed
crewai package (pip install crewai)
Active Vinkius subscription with a valid endpoint token

1

Install CrewAI
Run pip install crewai to install the framework. MCP support is built-in via the mcps parameter.
2

Add the MCP URL to your agent
Pass your Vinkius endpoint directly to the mcps list. Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com. CrewAI handles tool discovery and caching automatically.
3

Kick off your crew
Create a Crew with your agent and tasks. Call crew.kickoff() — the agent will automatically invoke Braintrust tools as needed.

crew.py

from crewai import Agent, Task, Crew

agent = Agent(
    role="Braintrust Analyst",
    goal="Access and analyze Braintrust data via MCP.",
    backstory="Expert analyst with direct Braintrust access.",
    mcps=[
        "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
    ],
)

task = Task(
    description="List recent Braintrust transactions",
    agent=agent,
    expected_output="A summary of recent activity",
)

crew = Crew(agents=[agent], tasks=[task])
result = crew.kickoff()
print(result)

Get your connection token →

Prerequisites

Python 3.10+ installed
crewai + crewai-tools packages
Active Vinkius subscription with a valid endpoint token

1

Install dependencies
Run pip install crewai crewai-tools. The MCPServerAdapter handles lifecycle management and tool conversion.
2

Connect with MCPServerAdapter
Use MCPServerAdapter as a context manager with SseServerParameters pointing to your Vinkius endpoint. The adapter automatically manages connection lifecycle.
3

Assign tools and run
Pass the returned mcp_tools to your agent's tools parameter. The adapter converts MCP tools to native BaseTool objects compatible with all CrewAI agents.

crew.py

from crewai import Agent, Task, Crew
from crewai_tools import MCPServerAdapter
from mcp import SseServerParameters

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

with MCPServerAdapter(server_params) as mcp_tools:
    agent = Agent(
        role="Braintrust Analyst",
        goal="Access and analyze Braintrust data via MCP.",
        backstory="Expert analyst with direct Braintrust access.",
        tools=mcp_tools,
    )

    task = Task(
        description="List recent Braintrust transactions",
        agent=agent,
        expected_output="A summary of recent activity",
    )

    crew = Crew(agents=[agent], tasks=[task])
    result = crew.kickoff()
    print(result)

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 Braintrust. 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 Braintrust 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 Braintrust MCP in CrewAI

Pass your Vinkius HTTP endpoint directly into the agent's `mcps` array parameter during initialization. CrewAI automatically fetches the tool schemas and exposes actions like `list_datasets` to that specific agent.

Yes, agents in the same crew share access to the configured MCP Server. While one agent queries `get_dataset`, another can write results using `create_experiment` within the same execution flow.

Use the `MCPServerHTTP` class from `crewai.mcp` and specify a `tool_filter`. This lets you restrict a junior agent to only run `list_prompts` while preventing it from modifying datasets.

Yes, it works natively over stdio and HTTP transports. You can run your crew locally in your terminal or deploy it to a headless CI/CD pipeline for automated regression testing.

Absolutely. Your prompt templates and dataset schemas are accessed via secure, point-to-point HTTPS connections. Vinkius handles the authentication token layer, ensuring your private prompt text never leaks into public model training logs or external storage.

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