- AUDIT
- COMPARE
- RECALL
- HUMAN
Short answer
How can I check my training parameters using AI?
Your agent uses Comet ML to fetch specific configuration data like learning rates or batch sizes. You get a direct readout of your experiment history through these queries.
The results you get.
Where your data lands.
The AI handles the retrieval so you can focus on the model logic.
AUDIT
Verify specific settings
The agent pulls exact values for any parameter you name. You get the precise numbers used in previous runs.
COMPARE
Spot configuration shifts
Ask the agent to look at multiple runs. It identifies how changes in batch size or learning rate affected your results.
RECALL
Find old experiments
Instead of scrolling through the UI, tell your agent to find a specific run by name. It finds the right experiment for you.
HUMAN
Decide next steps
The AI provides the raw data and summaries. You decide which hyperparameter ranges to test in your next training cycle.
The workflow
What your AI does when the request arrives.
The agent moves from finding the right experiment to pulling the exact data you asked for.
Locate experiments
The agent scans your Comet ML project to find the specific runs you are interested in.
list_experimentsExtract parameters
Once the run is identified, the agent pulls the specific configuration values like learning rate.
get_experiment_paramsSummarize findings
The agent organizes the retrieved parameters into a readable format for your review.
get_experiment_paramsReport to you
The agent delivers the final data directly in your chat interface.
list_experiments
Try it
Copy these to start.
Use these prompts to begin auditing your runs.
Starting points
Swap out the experiment names and parameter types with your actual Comet ML data.
Comet ML Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_C2PxqYqwffQi3V41WE8jlPynFSP048sA8SVR1PFT/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 — Comet ML capabilities are ready to use.
{
"mcpServers": {
"comet-ml-mcp": {
"url": "https://edge.vinkius.com/vk_preview_C2PxqYqwffQi3V41WE8jlPynFSP048sA8SVR1PFT/mcp"
}
}
}List my last five experiments in the 'ResNet-50' project.
What was the learning rate used in the experiment named 'train_run_v2'?
Compare the batch size between 'experiment_alpha' and 'experiment_beta'.
Find the optimizer settings for the most recent successful run.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Comet ML once, then ask.
One connection gives your agent access to your experiment history. Your credentials stay encrypted and secure.
Connect Comet ML to your AIFAQ
How this task behaves.
- 01
Can the AI change my hyperparameters?
No. The agent can only read and retrieve existing parameters from your experiments.
- 02
Can the AI delete an experiment?
No. This connector only has permission to list and get parameters; it cannot delete or modify data.
- 03
Does the AI see all my projects?
The agent can see any experiment it has permission to list via the connected account.
- 04
Can I ask for metrics like accuracy?
Yes, if those metrics are stored as part of the experiment parameters.
- 05
Do I need to keep the Comet ML tab open?
No. Once connected, the agent communicates with the API directly.
More questions about Comet ML? The Connector page answers them. See everything the Comet ML Connector can do
Connect Comet ML to Claude, Cursor, ChatGPT & more
