AI Model Evaluation and Testing for
Claude Code

AI Model Evaluation and Testing for
Claude Code


Route every AI API call from your Claude Code sessions through Neurometric for automatic logging, replay, and evaluation — with zero changes to your application code.



Route every AI API call from your Claude Code sessions through Neurometric for automatic logging, replay, and evaluation, with zero changes to your application code.


HOW TRAFFIC FLOWS
Your Code
Any AI SDK call
Neurometric Gateway
api.neurometric.ai
Logged · Evaluated · Stored
Upstream Provider
Anthropic · OpenAI · Mistral...

Every request, automatically

Every request, automatically

Neurometric intercepts traffic at the SDK layer, capturing rich metadata on every call.

Prompts & completions

Full input and output content for every request routed through the gateway.

Token usage

Input, output, and total token counts per request for cost tracking.

Model & Provider

Which model was called, on which provider, and with what parameters.

Latency & timestamps

Request timing data to help identify slow calls and performance regressions.

Up and running in 3 steps

No configuration files to edit. No code to change. Just clone, export, and go.

1. Clone the plugin into your Claude plugins directory

$ git clone https://github.com/NeurometricAI/neurometric-plugin \ ~/.claude/plugins/neurometric

2. Export your Neurometric API key

$ export NEUROMETRIC_API_KEY="sk_live_your-api-key"

3. Start Claude Code with the plugin enabled

$ claude --plugin-dir ~/.claude/plugins/neurometric


Heads up: Neurometric does not support stream: true requests. Make sure streaming is disabled in your SDK calls. Only models configured on your Neurometric account are accessible.

Two slash commands, right in Claude Code

Once the plugin is loaded, these commands are available from any Claude Code session.

/neurometric-status

Verifies that the Neurometric gateway is reachable, your API key is valid, and all SDK environment variables have been injected correctly.

/neurometric-replay [count]

Fetches and displays your most recent captured API calls directly in the terminal. Pass an optional count (default: 5) to control how many to show.

Learn more


AI-Powered Workspace

Ready to Deploy Intelligence Without Compromise?

Join enterprises that have eliminated API dependencies, slashed inference costs, and deployed AI that respects data sovereignty.