I'm building LiquidIndex, an RAG API service that makes it easy to add personalized context to AI apps. Instead of building a whole system to upload, process, and search through a user’s files or notes, you just call a few APIs. It handles the hard parts - like connecting to Notion or Google Drive, breaking the data into usable pieces, and making it searchable. Then, when you want to answer a question using that data, you just query it, and it gives you the most relevant results - optionally with an LLM-generated answer. No pipelines, no infrastructure, no headaches.

Here are the core APIs:

1. Create a customer (This is a space to put data)

2. Create an upload session (This is where users upload their data)

3. Query

Current connectors: File Upload, Google Drive, Notion, Dropbox

Supported File Types: PDF, text files, Markdown, CSVs, and XLSX (these include google docs and sheets)

Website: https://liquidindex.dev/

Check out the playground to get a feel of how it works!