Hero Light

What is Carbon?

Carbon provides a comprehensive framework designed to streamline the process of connecting external data sources to Large Language Models (LLMs).

Carbon simplifies the process of retrieval augmented generation (RAG), allowing you to spend more time using your data, and less time trying to ingest it.

How Carbon Works

Use our REST API or SDKs to sync user data sources and retrieve the data to use with LLMs. Carbon has native integrations with 20+ data connectors and supports more than 20+ file formats, encompassing text, audio, and visual data.

Depending on your use case and in-house infrastructure, you can retrieve user data from Carbon in several formats:

  • Parsed plaintext files
  • Embeddings (and chunks) to store in your vector store
  • Direct semantic and keyword search on a managed or self-hosted vector database


πŸ”— Connect

Client-side UI for users to authenticate and upload content from data sources such as Notion, Google Drive, Dropbox, OneDrive, websites, and file uploads. Available as a React component, SDK libraries, and soon as a magic link.

Carbon handles authentication (OAuth 1 & 2, API key-based, etc) for 20+ sources, processes the source data, and automates data synchronization.

πŸ—„οΈ Store

Choose between Carbon’s managed vector database (hosted on Qdrant Cloud) or your storage solution. The database updates as users modify connected sources and manage chunks alongside embeddings.

πŸ”Œ Universal API

Access and manage data (documents, chunks, vectors, etc.) from any source using our flexible API suite. Our API powers everything β€” our Carbon Connect UI makes calls to the backend.


πŸ”‘ Getting a Carbon API Key

Carbon is free to use up to the first million characters. Book a 15 minute onboarding to get an API key here.

To get started with Carbon, follow our guides: