Quickstart

Create a protected project and send traffic through Panicly

This guide follows the actual onboarding model from the web app: choose a plan, create a project, connect a provider, generate a Panicly key, then prove the setup with a first routed request.

Create or enter a workspace

Choose a plan

Free is enough to evaluate Panicly through OpenRouter. Direct OpenAI and Anthropic provider access require Pro.

Create a project

Pick a project name and environment. The project becomes the unit that owns API keys, provider key, rules, request logs, and model controls.

Connect a provider

Choose OpenRouter, OpenAI, or Anthropic. Store the provider key through the project settings flow.

Generate a Panicly API key

Copy the generated key immediately. It is returned once and then stored only as a prefix plus hash.

Send a first protected request

Route a provider-compatible request to Panicly. Onboarding watches for request evidence before considering the setup operational.

First request

curl

JavaScript

chat-completion.sh

$ curl https://panicly.vercel.app/v1/chat/completions \
>   -H "Authorization: Bearer $PANICLY_API_KEY" \
>   -H "Content-Type: application/json" \
>   -d '{
>     "model": "openrouter/auto",
>     "messages": [
>       { "role": "user", "content": "Say hello through Panicly." }
>     ],
>     "max_tokens": 64
>   }'

Operational, not just configured

The setup is complete only when Panicly sees traffic and can show request evidence in the dashboard.

Store keys intentionally

Panicly API keys are shown once. Provider keys are encrypted and used server-side for upstream forwarding.

$	curl https://panicly.vercel.app/v1/chat/completions \
>	-H "Authorization: Bearer $PANICLY_API_KEY" \
>	-H "Content-Type: application/json" \
>	-d '{
>	"model": "openrouter/auto",
>	"messages": [
>	{ "role": "user", "content": "Say hello through Panicly." }
>	],
>	"max_tokens": 64
>	}'