Workspace / Overview
Make every request count.
A clear place to manage access, understand usage, and reach the models behind your applications.
Demo mode is active.Add
INFERHUB_API_KEY to your Pages environment to connect live model data. The interface remains safe to explore.Start here
One key, familiar APIs.
Point OpenAI or Anthropic-compatible clients at Smart Key. We keep the provider credential server-side and give your team one surface to operate.
OPENAI_BASE_URL
https://smart-key-fh2.pages.dev/apiSetup
Workspace readiness
- Smart Key workspaceReady to configure
- InferHub connectionChecking provider status
- First API keyCreate a key when ready
Workspace / API keys
Keys that stay in your control.
Create client credentials for your applications. Secrets are shown once and never stored in the browser.
Your keys
Use separate keys for separate applications and rotate them without downtime.
Loading keys…
Workspace / Usage
See what your traffic is doing.
A practical request log for cost, latency, and model decisions.
Recent requests
Live records appear here when the provider is connected.
| Time | Model | Status | Tokens | Cost |
|---|---|---|---|---|
| Loading usage… | ||||
Workspace / Models
Know what is available.
Browse the provider catalogue and check the route before you ship it into an application.
Model catalogue
Availability and pricing are read from InferHub when configured.
Loading catalogue…
Learn / Docs
Ship with a familiar contract.
Smart Key keeps the integration close to the clients your team already knows.
OpenAI-compatible
Use the chat completions contract with any client that supports a custom base URL.
Base URL
https://smart-key-fh2.pages.dev/api/v1Header
Authorization: Bearer sk-smart-…Anthropic-compatible
Use the Messages API with the same Smart Key credential and a server-side provider connection.
Base URL
https://smart-key-fh2.pages.dev/apiHeader
x-api-key: sk-smart-…Build safely
Never put provider credentials in frontend code. Smart Key Functions keep INFERHUB_API_KEY in Cloudflare’s encrypted environment and forward only the request needed for inference.
Before productionConfigure an InferHub key, add your own auth/session layer, and set a custom domain. The current login is an honest MVP shell, not an account system.