Set up your API key and start working with Pinny

Pinny requires an eligible paid Layerpin plan and your own provider API key. Trial accounts must upgrade to send messages. Supported providers are OpenAI, Anthropic Claude, Google Gemini API, Google Vertex AI, xAI (Grok), Mistral AI, and DeepSeek. Your provider account needs API access and any required credits. A consumer chat subscription is not an API key.

Choose a provider you already use, or compare the approximate input and output rates in Pinny’s model picker. The API keys dialog links to each provider’s key page. Gemini API and Vertex use separate credentials: for Vertex, use an Express Mode API key, not a service-account file.

Open Pinny from the dashboard or map editor. On a narrow panel, header actions are grouped under More chat actions. Never paste your secret key into a message, a map field, or a screenshot.

Start with “How much plan capacity have I used?” in dashboard chat. For map work, open a map and ask “Suggest filters for these rooms” or attach a reference file. Image-based placement needs a model marked Images; text-only models can work with explicit coordinates. Check the proposal, apply it, and save the draft. Publish separately after previewing. The Pinny guide gives more ideas.

Keys are encrypted before storage and used by Layerpin’s server for your provider requests. Replace or remove them in API keys; revoke compromised keys with the provider. Saving a key does not test it: the provider checks it when you send a message. If a request fails, follow the specific key, model access, quota, or rate-limit message.

Use Chat history to return to conversations and Token usage to explore a date range and estimated provider costs. Deleting a chat does not delete its usage records. Provider invoices can differ because of caching, images, search, region, or discounts. Your Layerpin plan and provider API charges are separate.