Quickstart
This guide takes you from a fresh sign-in to your first auto-routed API call.
1. Sign in
Go to router.cocolevio.com. You land on the admin sign-in. Enter the workspace admin username and password (set by your administrator). There is no public sign-up — credentials are provisioned by the Cocolevio team.
2. Tour the sidebar
The left sidebar combines the platform pages with the Cocolevio additions:
| Sidebar item | What it does |
|---|---|
| Dashboard | Cost and usage overview — spend per day, per key, per model, with key management actions |
| Virtual Keys | Create and manage API keys, budgets, and expiries |
| Auto-Switch | Configure which models handle each complexity tier — changes apply live |
| AI Console | Chat workspace to test the router — shows the live routing log |
| Lead Agent | On-demand company research agent that writes enriched leads into Coco CRM |
| Playground | the raw model playground |
| Models + Endpoints | Every registered model and provider |
| Agentic → Agents | A2A agent registry (the Lead Agent is listed here with a ▶ Run button) |
| Usage / Logs | native usage analytics and per-request logs |
3. Create an API key
Open Virtual Keys → + Create New Key.
- Give it a Key Alias. If the alias starts with
auto(e.g.auto-myapp), the key becomes a universal auto-routing key: every request through it is auto-switched, no matter what model name the caller sends. - Optionally set a Max Budget (recommended for keys that ship inside apps) and an expiry.
- Copy the
sk-…value — it is shown only once.
4. Make your first call
Point any OpenAI-compatible client at the gateway:
curl https://router.cocolevio.com/dash/v1/chat/completions \
-H "Authorization: Bearer sk-YOUR-KEY" \
-H "Content-Type: application/json" \
-d '{"model":"auto","messages":[{"role":"user","content":"hi"}]}'
The response includes an ai_router field showing what the router decided:
"ai_router": {"tier": "SIMPLE", "model": "openrouter/meta-llama/llama-3.1-8b-instruct"}
Send a coding question and the same call routes to a strong coding model; send a proof or multi-step logic problem and it routes to a reasoning model. You never change your code.
5. Watch it work
Open the AI Console tab and send a few prompts of different difficulty — the live routing log on the right shows the tier, the model chosen, latency, tokens, and cost for every request. Then open Dashboard to see spend accumulating per key and per model.
Next steps
- Understand the tiers in Auto-Switch
- Change tier models live in the Control Panel
- Wire a real application in Integrating an application