powering real time finance data extraction
summary
CloudSLP is a syncing query engine built on DuckDB that pulls transactional data from many bank accounts (average person 2–4 accounts, businesses ~5) into a combined history and exposes reusable SQL templates via an API. A natural‑language front end uses an LLM to produce stepwise SQL which is tested sequentially against DuckDB WebAssembly and filtered for correctness. Exports include CSV, Excel, Google Sheets, APIs and a custom GPT action. DuckDB runs in browser, in Node.js preprocessing and in Python.
AI-written summary · about
chapters (9)
- 0:00The problem with bank data
- 1:00Introducing Cloudslurp
- 1:31UI and data export options
- 2:18Use cases and data tools
- 2:40LLM decision log for auditing
- 3:15Advanced workflows and ReAct
- 3:38DuckDB implementation
- 3:51DuckDB features and feedback
- 4:14Conclusion and Q&A
related talks
28mAnalyze data easily with duckdb - and the implications on data architectures
07mAerodynamic data models: Flying fast at scale with DuckDB
25mA quack of all trades: DuckDB, a versatile analytical database to keep in your toolkit
47mPlay Data Engineer Locally with DuckDB
55mDuckLake: Learning from Cloud Data Warehouses to Build a Robust “Lakehouse”
47mWhy DuckDB could be the engine for your Data Lake?
38mFrom Big Data to DuckDB: Ryan Boyd (MotherDuck) on Hyper-Tenancy, Duck Lake, and Making Data Fun.
27mDataEngBytes 2023 - MEL-T1-07 - Ryan Boyd