How to Use the Braintrust MCP in Windsurf

Q: How does Windsurf Cascade run a Braintrust experiment?

Cascade uses createexperiment to start a run, then pulls data using getdataset to feed the test cases. It handles the entire sequence autonomously without requiring individual prompts for each step.

Q: Can I manage Braintrust projects directly from the Windsurf UI?

Yes, you can list existing projects using listprojects or start fresh ones with createproject. The MCP panel in Windsurf displays these tools for easy access.

Q: How do I update Braintrust datasets using Windsurf?

Cascade invokes insertdatasetrow to add test targets. It extracts the raw code outputs from your editor and formats them to fit your schema.

Q: Can Cascade check my API keys?

The agent uses listenvvars to read your gateway setup. This ensures your Windsurf session uses the correct keys for model routing.

Run autonomous LLM evaluations and manage datasets directly inside Windsurf via this MCP Server.

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 Windsurf

Create your Vinkius account to connect Braintrust to Windsurf 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 Windsurf

Ask AI about this MCP

ChatGPT

Claude

Perplexity

Run Cascade evaluations via this MCP Server

The `create_experiment` tool lets Cascade spin up a new evaluation run to track your LLM outputs directly from the editor. Instead of jumping between tools, Windsurf chains this setup to log historical test runs the second you modify your prompt files. Cascade reads your local changes, triggers the test, and updates the score history. It keeps your development loop tight by keeping the evaluation results tied to your active branch.

Inject live test cases into Braintrust datasets

The `insert_dataset_row` tool appends new test cases directly into your evaluation matrices. Cascade detects failing edge cases in your Windsurf workspace and immediately pushes them to your gold-standard test suite. You don't have to manually format JSON files or copy-paste logs. Your agent pulls the raw execution data and updates the Braintrust matrix so your next evaluation run is ready with fresh data.

Pull versioned prompts straight to your workspace

The `get_prompt` tool pulls raw templates and variable contexts from your central registry. Cascade uses this to test prompt changes locally against your active Windsurf code before you commit them. This setup stops broken templates from hitting production. Your agent pulls the exact version you need and runs local validations without manual downloads.

Setup guide

Set up Braintrust MCP in Windsurf

Prerequisites

Windsurf IDE installed (macOS, Windows, or Linux)
Active Vinkius subscription with a valid endpoint token

1

Open MCP configuration

Click the Cascade assistant icon in the sidebar, then click the hammer icon (🔨) at the top of the panel. Select "Configure" to open ~/.codeium/windsurf/mcp_config.json.
2

Add the Braintrust MCP

Paste the JSON snippet shown on the right into the mcpServers object. Replace [YOUR_TOKEN_HERE] with your endpoint token from cloud.vinkius.com.
3

Refresh MCPs

Go back to the hammer icon (🔨) in Cascade and click "Refresh". Windsurf will detect the new server. No full restart is needed — the connection is hot-reloaded.
4

Verify in Cascade

Start a new Cascade conversation and ask something like "Show my Braintrust payment history." If connected, Cascade will call the Braintrust tools directly. You will see a green dot next to the server name in the MCP panel.

mcp_config.json

{
  "mcpServers": {
    "braintrust-mcp": {
      "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
    }
  }
}

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 Windsurf

Cascade uses `create_experiment` to start a run, then pulls data using `get_dataset` to feed the test cases. It handles the entire sequence autonomously without requiring individual prompts for each step.

Yes, you can list existing projects using `list_projects` or start fresh ones with `create_project`. The MCP panel in Windsurf displays these tools for easy access.

Cascade invokes `insert_dataset_row` to add test targets. It extracts the raw code outputs from your editor and formats them to fit your schema.

The agent uses `list_env_vars` to read your gateway setup. This ensures your Windsurf session uses the correct keys for model routing.

All prompt templates and API configurations remain isolated within Vinkius's secure sandbox. Windsurf accesses Braintrust via an encrypted token, meaning your raw credentials never touch local disk or external logs.

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