Full LLM Observability in 2 Minutes

Just Change Your Base URL

Keep your existing OpenAI/Anthropic credits. Get full observability with zero code changes. Debug latency issues, track costs, and never miss a failed request again.

  • Bring Your Own Keys – Use your provider accounts, pay them directly
  • 50+ fields per call – Costs by category, latency breakdown, full conversation history
  • Zero friction – Change one URL, instant visibility
Start Free (No Card Required) See the Code

100 free credits included. BYOK: 10% platform fee, you pay providers directly.

Before → After 2 lines changed 
# Before (standard OpenAI)
client = OpenAI(
    api_key="sk-..."
)

# After (with Demeterics observability)
client = OpenAI(
    base_url="https://api.demeterics.com/openai/v1",
    api_key="dmt_..."  # includes your OpenAI key
)

Works with Python, Node.js, Go, Rust, and any HTTP client

Zero Lock-In

Bring Your Own Keys (BYOK)

Keep your existing OpenAI, Anthropic, or Groq accounts. Route traffic through Demeterics, pay providers directly, and add a transparent 10% platform fee for observability.

Keep Your Credits Existing provider credits? Keep using them.
Transparent Pricing Provider rates + 10% platform fee. No hidden costs.
Switch Anytime Remove Demeterics? Just change back the URL.
Keep Your Rates Volume discounts and negotiated pricing stay intact.
Two Ways to BYOK
Option 1: Store Your Keys
Register once in Settings → API Keys. Encrypted with Google Cloud KMS.
Authorization: Bearer dmt_your_key
Option 2: Dual-Key (Per-Request)
Pass your vendor key with each request. Never stored.
Bearer dmt_your_key;gsk_vendor_key
Either way:
  • You pay providers directly
  • Full observability in your dashboard
  • 10% platform fee for the service
Game Changer

Tag Your Prompts. Zero Overhead.

Add metadata, A/B test variants, and audit notes directly in your prompts using /// NAME VALUE syntax. We strip these before the LLM sees them — so you pay nothing extra.

Works Everywhere No SDK, no headers — just edit your prompt
Tags Cost Nothing Stripped before reaching the LLM
A/B Testing Built-In Track variants without code changes
Perfect for No-Code Zapier/Make can't set headers, but CAN edit prompts
See Full Documentation
Your Prompt Tags stripped before LLM 
/// APP customer-support
/// FLOW ticket-response
/// VARIANT gpt4-concise
/// VERSION 2.3.1

You are a helpful customer support agent.

/// This variant uses shorter responses
/// Approved by legal team 2024-12-15

Respond to this ticket: {{{ ticket_text }}}
Metadata Flows to BigQuery
Comments Stripped, not billed
Variables {{{ }}} for templates

Everything You Need. Nothing You Don't.

From debugging to ROI tracking — features that actually matter.

Latency p95
Debug slow calls
Error Tracking
Rate limits, 500s
Cost Attribution
By team/workflow
/// Tagging
Free metadata tags
Dual-Key Auth
Secure automation
BYOK
Your keys + 10% fee
Council API
18 persona feedback
Conversion Track
Link AI to ROI
A/B Testing
Prompt variants
Data Export
CSV via API
Budget Guards
Per-key limits
ZDR Mode
Control what's stored

See It In Action

From black box to glass box: watch how Demeterics turns AI spending into crystal-clear visibility.

Full demo: n8n workflows, cost tracking, multi-agent quality control

Start Free Now Read the Docs
Zero Friction

Integration = Change One URL

No SDK to install. No code rewrite. Just change your base URL and you're done. 

# pip install openai
from openai import OpenAI

# Before
# client = OpenAI(api_key="sk-...")

# After (with Demeterics observability)
client = OpenAI(
    base_url="https://api.demeterics.com/openai/v1",
    api_key="dmt_your_key"  # includes your OpenAI key
)

response = client.chat.completions.create(
    model="gpt-4",
    messages=[{
        "role": "user",
        "content": """/// APP my-app
/// FLOW chat
Hello!"""
    }]
)
// npm install openai
import OpenAI from 'openai';

// Before
// const client = new OpenAI({ apiKey: "sk-..." });

// After (with Demeterics observability)
const client = new OpenAI({
    baseURL: "https://api.demeterics.com/openai/v1",
    apiKey: "dmt_your_key"
});

const response = await client.chat.completions.create({
    model: "gpt-4",
    messages: [{
        role: "user",
        content: `/// APP my-app
/// FLOW chat
Hello!`
    }]
});
# Just change the URL - that's it!
curl https://api.demeterics.com/openai/v1/chat/completions \
  -H "Authorization: Bearer dmt_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4",
    "messages": [{
      "role": "user",
      "content": "/// APP my-app\n/// FLOW chat\nHello!"
    }]
  }'
# pip install langchain-openai
from langchain_openai import ChatOpenAI

# Before
# llm = ChatOpenAI(model="gpt-4", api_key="sk-...")

# After (with Demeterics observability)
llm = ChatOpenAI(
    model="gpt-4",
    base_url="https://api.demeterics.com/openai/v1",
    api_key="dmt_your_key"
)

# Tags go in your prompt - stripped before LLM, zero cost
response = llm.invoke("""/// APP langchain-app
/// FLOW rag-pipeline
Hello!""")
Quick Start Guide Get Your API Key

Works With All Major LLM Providers

One proxy. One dashboard. All your AI.

🤖
OpenAI
🧠
Anthropic
âš¡
Groq
💎
Google
🔀
OpenRouter

Integrations

Works With Your Existing Stack

LangChain
Vercel AI
Zapier
Make
n8n
Any HTTP

Start Free. Scale When Ready.

100 free credits to start. BYOK: pay providers directly + 10% platform fee.

100 Free Credits No card required
BYOK Mode Your keys + 10% fee
Unlimited Users No per-seat fees
Start Free Now View Full Pricing

Ready to See What's Happening in Your LLM Stack?

Change one URL. Get instant visibility. Debug faster. Spend smarter.

Start Free (2 Minutes to Setup) See the Code

100 free credits. BYOK supported. No credit card required.