- DATA
- WEIGHTS
- LOGS
- HUMAN
Short answer
How can I audit model artifacts with my AI?
Your AI agent uses Weights & Biases to locate specific versions of datasets and model weights. It pulls the metadata you need to verify the exact files used in a research run.
Audit outcomes.
Where your data lands.
The AI handles the search, leaving you with the specific files you need.
DATA
Version tracking
The AI identifies the exact dataset version tied to a specific run. You get the precise file path or identifier.
WEIGHTS
Model weight retrieval
Your agent finds the specific weights used during training. This ensures you aren't guessing which checkpoint is which.
LOGS
Artifact metadata
The AI pulls the metadata associated with an artifact. You see the size, type, and creation timestamp for every file.
HUMAN
Manual verification
The AI provides the list of files, but you decide which ones to download or use for your next training run.
The workflow
What your AI does when the request arrives.
The agent acts as a bridge between your chat interface and your experiment logs.
Identify the project
The agent looks up the target project to narrow down the search space for artifacts.
list_project_artifactsScan for artifacts
The AI crawls the project to find all registered datasets, models, and weights.
list_project_artifactsFilter by version
Your agent compares the requested run ID against the available artifact versions to find a match.
list_project_artifactsReport findings
The AI presents a clean list of the specific files, versions, and metadata you asked for.
list_project_artifacts
Try it
Copy these to start.
Use these prompts to begin your audit.
Starting points
Swap out the project names and run IDs with your actual Weights & Biases identifiers.
Weights & Biases Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_I1Yu40vJsqaxsdrmMLW0Zf84kGzji4jYFvnrqvBH/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Weights & Biases capabilities are ready to use.
{
"mcpServers": {
"weights-biases-mcp": {
"url": "https://edge.vinkius.com/vk_preview_I1Yu40vJsqaxsdrmMLW0Zf84kGzji4jYFvnrqvBH/mcp"
}
}
}List all the model weight artifacts in my 'transformer-v2' project.
Which version of the 'imagenet-subset' dataset was used in the last run of the 'resnet-training' project?
Find the specific model weights used in the run with ID 'abc123def'.
Show me all artifacts in the 'nlp-research' project that were created last Tuesday.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Weights & Biases once, then ask.
Just click the link to connect your account. Your credentials stay encrypted, and you can start asking questions immediately in your preferred AI client.
Connect Weights & Biases to your AIFAQ
How this task behaves.
- 01
Can the AI delete artifacts from my project?
No. The AI can only read and list artifacts. It cannot delete or modify any data in Weights & Biases.
- 02
Can the AI download the weights directly to my local machine?
No. The AI retrieves the metadata and file paths, but you must handle the actual file downloads yourself.
- 03
Does the AI need my Weights & Biases API key?
Yes, but you provide it through the secure Vinkius connection, not directly to the AI model.
- 04
Can I ask about artifacts in a different project?
Yes, as long as you specify the project name in your prompt.
- 05
Can the AI create new artifacts?
No. This specific capability is limited to identifying and retrieving existing artifacts.
More questions about Weights & Biases? The Connector page answers them. See everything the Weights & Biases Connector can do
Connect Weights & Biases to Claude, Cursor, ChatGPT & more
