GitHub hosts HunyuanVideo, Tencent's open-source framework for large-scale video generation models, enabling AI-driven video creation.
White-Label RAG API
0Popularity
About White-Label RAG API
The White-Label RAG API provisions a dedicated vector index for each customer, built from their website or custom data feeds. The index is created with a single API call at signup and deleted at churn, ensuring customer data remains isolated without requiring the provider to host a vector database. Content can be sourced from crawled websites or pushed via the Custom Connections API, with markdown formatting applied automatically. The system handles chunking, embeddings, and scoring, using standard OpenAI embeddings and a blended algorithm that weights passages by length, source relevance, and document structure. Endpoints return either raw context passages or concise cited answers, enabling personalized AI responses grounded in each customer’s own content. The API also supports organization-level content that blends into every customer’s index, reducing duplication. Admin tools include an MCP server for provisioning, inspecting accounts, and querying indexes directly from AI assistants.
Key features
- Provision customer indexes with one API call
- Crawl and chunk customer websites automatically
- Push custom markdown content via Custom Connections API
- Retrieve context passages or cited answers via API
- Blend organization-level content into all indexes
- Admin MCP for provisioning and fleet management
- Standard OpenAI embeddings with blended scoring
- Supports idempotent creation and deletion of indexes
Use cases
- Personalize AI responses for each customer using their own content
- Ground AI answers in customer-specific documentation or websites
- Automate onboarding with customer context pre-loaded
Pros
- One private vector index per customer
- No vector database or scoring algorithm to host
- Supports website crawling and custom markdown content
- Idempotent provisioning and deletion via API
- Organization-level content blends into all customer indexes
Cons
- No free tier mentioned
- Requires integration via API calls
- Limited to markdown-formatted content
- No explicit support for non-English content
White-Label RAG API videos
Frequently asked questions about White-Label RAG API
What is the White-Label RAG API and what does it do?
The White-Label RAG API provisions a private vector index for each customer, built from their website or custom data feeds. It enables personalized AI responses grounded in each customer’s own content by handling chunking, embeddings, and scoring automatically.
Who is the White-Label RAG API designed for?
The tool is designed for SaaS products and businesses that want to personalize AI responses, onboarding, or UI for each customer using their own content. It suits companies that need to avoid hosting vector databases or managing complex indexing pipelines.
How do I get started with the White-Label RAG API?
Start by provisioning a customer with a single API call at signup, including their customer ID and website URL if available. The system handles the rest, including indexing and provisioning, and you can manage accounts via the admin MCP or API.
Does the White-Label RAG API support custom data beyond websites?
Yes, the API supports pushing custom data via the Custom Connections API, including markdown-formatted content like dashboards, reports, or help articles. This data is stored, chunked, and indexed automatically without requiring you to host a database.
Can the White-Label RAG API blend content across multiple customers?
Yes, organization-level content, such as help docs or glossaries, can be added once and blended into every customer’s index automatically. This reduces duplication and ensures consistency across all customer indexes.
What endpoints does the White-Label RAG API provide?
The API provides endpoints for provisioning customers, retrieving context passages, and generating concise cited answers. It also includes an admin MCP for managing accounts, inspecting indexes, and pushing content directly from AI assistants.
