NeuraMeter
Stop wasting tokens on context your agents don’t need.
NeuraMeter is an open-source context engineering platform for AI agents. Analyze context efficiency. Set guardrails. Let agents self-optimize.
Why NeuraMeter?
Existing tools record costs. NeuraMeter optimizes them.
- Context efficiency analysis — See which agents waste tokens on stale context, oversized tool results, or bloated system prompts
- 3-mode guardrails — Choose between
notify(alert only),block(hard stop), orauto-optimize(self-heal and retry) - MCP server — Agents check their own cost and context in real-time, then self-optimize
- MIT License — Permanently open source
3-Layer Architecture
┌─────────────────────────────────────────────────────────┐
│ Layer 1: Measure │
│ Cost tracking · Context utilization · Dashboard │
├─────────────────────────────────────────────────────────┤
│ Layer 2: Guard │
│ notify (alert only) · block (hard stop) · │
│ auto-optimize (self-heal) │
├─────────────────────────────────────────────────────────┤
│ Layer 3: Optimize │
│ MCP server · Agents self-check cost & context · │
│ Autonomous optimization │
└─────────────────────────────────────────────────────────┘Get Started
npm install @neurameter/coreimport OpenAI from 'openai';
import { withNeuraMeter } from '@neurameter/core';
const openai = withNeuraMeter(new OpenAI(), {
apiKey: 'nm_xxx',
projectId: 'proj_xxx',
agentName: 'MyAgent',
});
// Just call it normally. Cost & context tracked automatically.
const response = await openai.chat.completions.create({
model: 'gpt-4o',
messages: [{ role: 'user', content: 'Hello' }],
});