Quick start

Integrate CoreCtic AI in under 3 minutes. No change to your application logic — a single URL swap is enough.

Full OpenAI compatibility

CoreCtic AI implements the OpenAI v1 spec at 100%. Any code that works with the official OpenAI SDK works with CoreCtic AI without any other change.

Coding with an AI? Let it integrate CoreCtic for you.

Give this file as context to Claude Code, Cursor or ChatGPT — it contains everything the assistant needs to know to hook your application up to CoreCtic.

corectic-llm-guide.md

Setup in 3 steps

Create your account & grab your API key

Go to corecticai.com/auth and create your account. An API key is generated automatically — it is always prefixed pl_.

You will find it in Dashboard → Settings → API Keys. Copy it and store it in an environment variable.

bash

# Add to your .env or .env.local
CORECTIC_API_KEY="pl_xxxxxxxxxxxxxxxxxxxx"

Replace the base URL

Only one parameter changes. Your api_key becomes your CoreCtic key (pl_...) and your base_url points to our proxy:

Python (before / after)

# Before — direct OpenAI call
from openai import OpenAI
client = OpenAI(api_key="sk-proj-...")

# After — via CoreCtic AI
from openai import OpenAI
import os

client = OpenAI(
    api_key=os.environ["CORECTIC_API_KEY"],   # pl_...
    base_url="https://api.corecticai.com/v1",
)

Node.js (before / after)

// Before
import OpenAI from 'openai';
const client = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });

// After
import OpenAI from 'openai';
const client = new OpenAI({
  apiKey: process.env.CORECTIC_API_KEY,   // pl_...
  baseURL: 'https://api.corecticai.com/v1',
});

Send your first request

All your existing code works as-is. The request is proxied, logged and optimized by the semantic cache:

Python

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Explain semantic caching in 2 sentences."}]
)
print(response.choices[0].message.content)

# The X-Cache response header indicates:
#   HIT  → served from the cache (0 tokens billed)
#   MISS → API call made (tokens counted)

Verify the integration

After your first request, open the Dashboard. You should see:

The request appear in Request logs with model, tokens and cost
The cache status switch to HIT from the second similar request
P50/P95 latency metrics updated in real time

Common troubleshooting

Error	Cause	Solution
401 Unauthorized	Invalid or missing API key	Check the key starts with pl_ and is correctly passed as api_key
403 Forbidden	Plan too low for this provider	Check your plan restrictions in Dashboard → Plans
422 Unprocessable	Unknown model	Check the exact model spelling (see Providers section)
429 Too Many Requests	Rate limit reached	Standard plan: 60 req/min. Pro: 300 req/min. Premium: unlimited
502 Bad Gateway	The AI provider is failing	Retry or switch models. CoreCtic AI logs all upstream errors.

Next steps

API integration

Streaming, env variables, custom headers

Semantic cache

Understand and monitor the cache

Providers

All supported models

Analytics

Metrics, costs and PDF export