Learn how to integrate, configure, and make the most of SupportPilot.
SupportPilot is an AI-powered customer support platform that trains on your documentation to answer customer questions instantly. It combines Retrieval-Augmented Generation (RAG) with customizable chat widgets to provide accurate, source-cited answers 24/7.
Getting started takes 5 minutes: (1) Create an account, (2) Upload your documentation, (3) Configure your AI settings, (4) Embed the widget on your site. The AI will begin answering questions immediately using your knowledge base.
Documents are uploaded files that the system chunks and indexes. Conversations store the history of customer interactions. The Widget is the embeddable chat interface. AI Config controls how your assistant behaves and responds.
The simplest way to add SupportPilot to your website is by inserting a script tag in your HTML. Copy the widget script from your dashboard settings and paste it just before the closing </body> tag. The widget will appear in the position you configured.
From the dashboard widget builder, you can customize the primary color, border radius, position (bottom-right, bottom-left), greeting message, launcher icon, and suggested questions. All changes preview in real-time before you deploy.
For React-based projects, use the supportpilot-react package. Install with npm, import the SupportPilotWidget component, and pass your configuration as props. The component works with both client-side and server-side rendered frameworks.
If the widget doesn't appear, check that (1) the script tag is correctly placed, (2) no ad blockers are interfering, (3) the company slug is correct, and (4) the widget is enabled in your dashboard settings.
All API requests require an API key passed via the Authorization header. Generate API keys from your dashboard settings. Keys are prefixed with sk_live_ and should be kept secret. Include the key as: Authorization: Bearer sk_live_...
POST /api/chat — Send a customer message and receive an AI-generated response. Supports streaming via Server-Sent Events. Request body includes conversationId, message, and optional companyId. Returns streamed response with source citations.
GET /api/documents — List all documents. POST /api/upload — Upload a new file. DELETE /api/documents/[id] — Remove a document. The system automatically processes uploaded documents through the chunking and embedding pipeline.
GET /api/conversations — List conversations with filters. GET /api/conversations/[id] — Get conversation detail with messages. PATCH /api/conversations/[id]/status — Update conversation status. POST /api/feedback — Submit customer feedback.
The JavaScript SDK provides a programmatic interface to the SupportPilot API. Install via npm: npm install supportpilot-js. Supports all API endpoints, including streaming chat, document management, and analytics queries.
For Python projects, install the supportpilot-py package: pip install supportpilot-py. The SDK mirrors the REST API with typed methods for every endpoint. Includes async support for high-throughput applications.
The widget exposes a JavaScript API for advanced integrations. Use window.SupportPilot to programmatically open/close the widget, trigger a new conversation, or listen for events like message received and conversation resolved.
When you upload a document, it's split into chunks, converted into vector embeddings, and stored in our vector database. When a customer asks a question, we find the most relevant chunks, inject them as context, and the AI generates an answer based on that context.
No. Your documents and conversations are never used to train or improve the underlying AI models. Each company's data is isolated at the row level, and we never share data between customers.
Yes. You can export all your data at any time, including documents, conversations, and analytics. Use the dashboard export feature or the API to download your data in standard formats.