#OpenSourceDiscovery 96: WrenAI
SQL AI Agent to chat with database and generate charts, reports, and BI
Today I discovered…
WrenAI
A toolchain consisting UI, AI Service, and Semantic Engine for data modelling, SQL generation using RAG architecture leveraging LLMs, and data visualisation
💖 What I like about WrenAI:
End-to-end solution: While having a modular structure, it has bootstrapping scripts and docker containers that put together various services to deliver BI directly from the data warehouse, handy for exploratory work
Supports almost all popular data warehouses including BigQuery, Snowflake, Postgres, etc.
Having natural language interface to the data helps think on the next level
👎 What I dislike about WrenAI:
It was unusable with local LlaMa models (served using Ollama). This was my primary motive to use WrenAI in the first place. Neither the quality was good nor the speed was within practical threshold where it could prove to be usable in daily solution.
Even using OpenAI and Anthropic models, it was pretty slow to respond on a top end computer (CPU only)
Did not work well with the JSON data schema. I wish for better support for unstructured data.
Overall, I like the direction where it is going. I will wait to use this project until the focus is on performance before features, this might require big engineering decisions including the overhaul of external dependencies.
⭐ Ratings and metrics
Based on my experience, I would rate this project as following
Production readiness: 7/10
Docs rating: 7/10
Time to POC(proof of concept): less than a day
Author: Chih-Yu Yeh, Andy Yen, William chang, Shimin, Pao Sheng, Freda Lai, Howard Chi, and other Canner team members
Demo | Source
🛡 License: AGPL-3.0
Tech Stack: Typescript, PythonNote: In my trials, I always build the project from the source code to make sure that I test what I see on GitHub. Not the docker build, not the hosted version.
If you discovered an interesting Open-Source project and want me to feature it in the newsletter, get in touch with me.
To support this newsletter and Open-Source authors, follow #OpenSourceDiscovery on LinkedIn and Twitter
Hi, thanks for posting this and share with others your reviews and perspective on Wren AI.
I am one of core members of Wren AI. Regarding to your "dislike about Wren AI", may I ask you two follow-up questions here:
1. Even using OpenAI and Anthropic models, it was pretty slow to respond on a top end computer (CPU only) -> Could you elaborate more on this? And what's the expected latency for you between asking a question and getting response back?
2. Did not work well with the JSON data schema. I wish for better support for unstructured data. -> I would also want to hear your use cases here.