| Model Requested | Routed To | Saved | Status |
|---|---|---|---|
| No requests yet. Make your first API call! | |||
All Clear
No alerts. Your gateway is running smoothly.
| Time | Requested | Routed To | Provider | Input | Output | Saved | Savings% | Status |
|---|---|---|---|---|---|---|---|---|
| No API requests yet. | ||||||||
Gateway API Keys
Use these keys with our OpenAI-compatible endpoint
Drop-in replacement for OpenAI. Just change base_url — nothing else.
| Name | Key | Requests | Last Used | Status | Actions |
|---|---|---|---|---|---|
| No keys yet. Create your first gateway key above. | |||||
🔗 BYOK — Bring Your Own Keys
Your provider keys never leave our encrypted vault. We route through them on your behalf.
AIMLAPI Key Get free key →
Sign up at aimlapi.com — free tier gives you 50K tokens/day. 400+ models.
CometAPI Key Get free key →
Sign up at cometapi.com — 620+ models, 20% cheaper on frontier tokens.
Verify & Activate
We'll test both keys and activate your gateway. Takes about 5 seconds.
📄 Quick Start
From zero to first routed request in under 5 minutes.
from openai import OpenAI # Just change base_url — everything else stays the same client = OpenAI( api_key="YOUR_COSTIMPLODE_KEY", base_url="https://api.costimplodeai.com/v1" ) response = client.chat.completions.create( model="gpt-4o", # we route to cheapest equivalent messages=[{"role": "user", "content": "Hello!"}] ) # Check what we routed to and how much you saved: # response.headers['x-ci-routed-model'] # response.headers['x-ci-saved-usd'] # response.headers['x-ci-savings-pct'] print(response.choices[0].message.content)
| Header | Value Example | Description |
|---|---|---|
| x-ci-routed-model | deepseek-v3 | What model actually handled the request |
| x-ci-savings-pct | 94 | Percentage saved vs. GPT-4o baseline |
| x-ci-saved-usd | 0.0048 | Exact USD saved on this request |
| x-ci-provider | aimlapi | Which provider was used |
| x-ci-latency-ms | 312 | Gateway latency added (typically <50ms) |
🤖 AI Support Agents
7 specialized agents. 20 languages. Mem0 + Redis memory. Running 24/7.
🛡️ Escalation Tickets
Only surfaces when agents can't resolve. Clean briefs only.
All Clear
No escalations. Agents are handling everything.
💳 Plan & Billing
BYOK model — your provider costs stay with you. We just charge for the gateway.