Spice AI
@spice_ai
Followers
4K
Following
259
Media
151
Statuses
484
Open-source data & AI platform that combines SQL query federation & acceleration, hybrid search, and LLM inference in a high-performance, lightweight runtime.
Seattle, USA
Joined December 2020
Your query engine doesn’t have to be a monolith. Spice is a 140 MB binary, so you can run it anywhere - including as a sidecar right next to your app for: - Ultra-low latency - Simple setup - Local acceleration per pod More details & deployment options: https://t.co/IQPQcB7JXK
0
0
1
Enterprise apps need to pull in data from many sources, but this historically comes with a latency tax. Spice lets you join that data and accelerate locally: in this example, the same query drops from 1.07 seconds to 0.011 seconds. Full cookbook here: https://t.co/571FUGCxSl
0
0
1
Want to add AI to your stack without new APIs or pipelines? With Spice’s ai() SQL function, you can run LLM calls right inside SQL. Classify sentiment, translate text, and combine results across datasets in one query. Check out the docs for more info: https://t.co/QIEEBDN4lD
0
0
0
Effective search means accounting for every factor that impacts relevance - like giving higher weight to recency. Explore how RRF improves relevance across multiple hybrid search use cases and learn how to start using it: https://t.co/MFUTKwt4LA
0
0
1
Meet the Spice AI team at #SeattleAIWeek! Find Spice in Startup Alley at the Seattle AI Week Summit all day today, and @0xLukeKim floating around the city speaking on various panels. Thanks to @WTIA for organizing this great week of events.
1
1
5
Spice v1.8 added support for writing to @ApacheIceberg tables with standard SQL - extending federation in Spice beyond reads to handle ingestion and transformation in the same runtime. Check out the full demo: https://t.co/TPcoKHtQSX, & the blog for the guide:
0
0
3
Multiple search modalities all ranked together in one SQL query with Spice using Reciprocal Rank Fusion (RFF). 🎥 Watch the full demo: https://t.co/sQzH9pR0dC 📘 Read the step-by-step guide: https://t.co/l0YqkUn4z8
0
1
3
With Acceleration Snapshots in v1.8, developers can now reuse pre-built accelerations from S3 instead of reloading from a federated source each time. @leblancphill demos how it works: ✅ Faster startup ✅ Less load on upstream systems ✅ Keeps apps online even if sources go
0
0
2
Spice is proud to sponsor #SeattleAIWeek! ⚡ Join us on Oct 28 in Startup Alley at the Summit to see how we’re powering real-time, AI-driven data apps. Thanks to the @WTIA team for organizing! Learn more and register here: https://t.co/SNoUnHnF1D
0
0
2
Developers use Spice + @awscloud S3 Vectors for scalable semantic search, recommendations, and RAG. All powered by SQL. Now with Partitioned S3 Vector Indexes, you can scale vector search efficiently across massive embedding sets. Learn more in the docs: https://t.co/8hbc68PODw
0
0
1