The agent can generate PDF documents from HTML or URLs, and images in PNG, JPEG, or WebP formats. For PDFs, you control paper size (letter, legal, A4), orientation, margins, headers, and footers. For images, you set dimensions, device scale factor, and transparency. Template-based PDFs also support Nunjucks templating with dynamic data injection.

The agent collects relevant details during the conversation, such as customer name, order number, or account balance. It passes this data as a JSON payload to Cloudlayer's Template-to-PDF endpoint along with your template ID. Cloudlayer merges the data into the Nunjucks template and returns the finished PDF.

No. Documents are generated and stored by Cloudlayer, either in their default storage or in your configured external bucket on AWS, Azure, or GCP. Tars retrieves the download URL from Cloudlayer and shares it with the customer. No copies are retained in Tars infrastructure.

Yes. Cloudlayer supports async generation with webhook callbacks. The agent submits the job, monitors its status through the Get Job endpoint, and notifies the customer once the document is ready. For simpler documents, synchronous generation returns the file immediately within the conversation.

Tars requires your Cloudlayer API key, which you can generate from the API Keys section in your Cloudlayer account settings. The key grants access to document generation endpoints. You can revoke or rotate the key anytime from your Cloudlayer dashboard without affecting other integrations.

Yes. Cloudlayer's URL-to-Image and URL-to-PDF endpoints support HTTP basic authentication and custom cookies. The agent can pass credentials and session cookies so the renderer accesses authenticated pages, internal dashboards, or gated content before capturing the output.

Manual attachment requires someone to create the document first, then upload it. With Tars and Cloudlayer, the AI agent generates the document on demand during the conversation using live data and templates. No human creates, saves, or uploads anything. The customer gets a fresh, accurate document in real time.

Yes. Cloudlayer supports external storage configurations for AWS S3, Azure Blob, and Google Cloud Storage. Through the Add Storage endpoint, you specify your bucket name, region, and credentials. All generated documents route to your own infrastructure, giving you full control over retention and access policies.