Spice is a portable runtime written in Rust that offers developers a unified SQL interface to materialize, accelerate, and query data from any database, data warehouse, or data lake.
📣 Read the Spice.ai OSS announcement blog post.
Spice connects, fuses, and delivers data to applications, machine-learning models, and AI-backends, functioning as an application-specific, tier-optimized Database CDN.
Spice is built-with industry leading technologies such as Apache DataFusion, Apache Arrow, Apache Arrow Flight, SQLite, and DuckDB.
🎓 Read the MaterializedView interview on Spice.ai
🎥 Watch the CMU Databases Accelerating Data and AI with Spice.ai Open-Source
Spice makes it fast and easy to query data from one or more sources using SQL. You can co-locate a managed dataset with your application or machine learning model, and accelerate it with Arrow in-memory, SQLite/DuckDB, or with attached PostgreSQL for fast, high-concurrency, low-latency queries. Accelerated engines give you flexibility and control over query cost and performance.
-
Application-focused: Spice is designed to integrate at the application level; 1:1 or 1:N application to Spice mapping, whereas most other data systems are designed for multiple applications to share a single database or data warehouse. It's not uncommon to have many Spice instances, even down to one for each tenant or customer.
-
Dual-Engine Acceleration: Spice supports both OLAP (Arrow/DuckDB) and OLTP (SQLite/PostgreSQL) databases at the dataset level, unlike other systems that only support one type.
-
Separation of Materialization and Storage/Compute: Spice separates storage and compute, allowing you to keep data close to its source and bring a materialized working set next to your application, dashboard, or data/ML pipeline.
-
Edge to Cloud Native. Spice is designed to be deployed anywhere, from a standalone instance to a Kubernetes container sidecar, microservice, or cluster at the Edge/POP, On-Prem, or in public clouds. You can also chain Spice instances and deploy them across multiple infrastructure tiers.
Spice | Trino/Presto | Dremio | Clickhouse | |
---|---|---|---|---|
Primary Use-Case | Data & AI Applications | Big Data Analytics | Interactive Analytics | Real-Time Analytics |
Typical Deployment | Colocated with application | Cloud Cluster | Cloud Cluster | On-Prem/Cloud Cluster |
Application-to-Data System | One-to-One/Many | Many-to-One | Many-to-One | Many-to-One |
Query Federation | Native with query push-down | Supported with push-down | Supported with limited push-down | Limited |
Materialization | Arrow/SQLite/DuckDB/PostgreSQL | Intermediate Storage | Reflections (Iceberg) | Views & MergeTree |
Query Result Caching | Supported | Supported | Supported | Supported |
Typical Configuration | Single-Binary/Sidecar/Microservice | Coodinator+Executor w/ Zookeeper | Coodinator+Executor w/ Zookeeper | Clickhouse Keeper+Nodes |
1. Faster applications and frontends. Accelerate and co-locate datasets with applications and frontends, to serve more concurrent queries and users with faster page loads and data updates. Try the CQRS sample app
2. Faster dashboards, analytics, and BI. Faster, more responsive dashboards without massive compute costs. Watch the Apache Superset demo
3. Faster data pipelines, machine learning training and inferencing. Co-locate datasets in pipelines where the data is needed to minimize data-movement and improve query performance. Predict hard drive failure with the SMART data demo
4. Easily query many data sources. Federated SQL query across databases, data warehouses, and data lakes using Data Connectors.
-
Is Spice a cache? No, however you can think of Spice data materialization like an active cache or data prefetcher. A cache would fetch data on a cache-miss while Spice prefetches and materializes filtered data on an interval or as new data becomes available. In addition to materialization Spice supports results caching.
-
Is Spice a CDN for databases? Yes, you can think of Spice like a CDN for different data sources. Using CDN concepts, Spice enables you to ship (load) a working set of your database (or data lake, or data warehouse) where it's most frequently accessed, like from a data application or for AI-inference.
-
Where is the AI? Spice provides a unified API for both data and AI/ML with a high-performance bus between the two. However, because the first step in AI-readiness is data-readiness, the Getting Started content is focused on data. Spice has endpoints and APIs for model deployment and inference including LLMs, accelerated embeddings, and an AI-gateway for providers like OpenAI and Anthropic. Read more about the vision to enable development of intelligent AI-driven applications.
BI.dashboard.acceleration.with.Spice.mp4
Currently supported data connectors for upstream datasets. More coming soon.
Name | Description | Status | Protocol/Format |
---|---|---|---|
duckdb |
DuckDB | Release Candidate | |
github |
GitHub | Release Candidate | |
graphql |
GraphQL | Release Candidate | JSON |
mysql |
MySQL | Release Candidate | |
postgres |
PostgreSQL | Release Candidate | |
s3 |
S3 | Release Candidate | Parquet, CSV |
databricks (mode: delta_lake) |
Databricks | Release Candidate | S3/Delta Lake |
databricks (mode: spark_connect) |
Databricks | Beta | Spark Connect |
delta_lake |
Delta Lake | Beta | Delta Lake |
file |
File | Beta | Parquet, CSV |
flightsql |
FlightSQL | Beta | Arrow Flight SQL |
odbc |
ODBC | Beta | ODBC |
spiceai |
Spice.ai | Beta | Arrow Flight |
abfs |
Azure BlobFS | Alpha | Parquet, CSV |
clickhouse |
Clickhouse | Alpha | |
debezium |
Debezium CDC | Alpha | Kafka + JSON |
dremio |
Dremio | Alpha | Arrow Flight |
ftp , sftp |
FTP/SFTP | Alpha | Parquet, CSV |
http , https |
HTTP(s) | Alpha | Parquet, CSV |
localpod |
Local dataset replication | Alpha | |
mssql |
Microsoft SQL Server | Alpha | Tabular Data Stream (TDS) |
sharepoint |
Microsoft SharePoint | Alpha | Unstructured UTF-8 documents |
snowflake |
Snowflake | Alpha | Arrow |
spark |
Spark | Alpha | Spark Connect |
Currently supported data stores for local materialization/acceleration. More coming soon.
Name | Description | Status | Engine Modes |
---|---|---|---|
arrow |
In-Memory Arrow Records | Release Candidate | memory |
duckdb |
Embedded DuckDB | Release Candidate | memory , file |
postgres |
Attached PostgreSQL | Beta | |
sqlite |
Embedded SQLite | Beta | memory , file |
quickstart.mp4
Step 1. Install the Spice CLI:
On macOS, Linux, and WSL:
curl https://install.spiceai.org | /bin/bash
Or using brew
:
brew install spiceai/spiceai/spice
On Windows:
curl -L "https://install.spiceai.org/Install.ps1" -o Install.ps1 && PowerShell -ExecutionPolicy Bypass -File ./Install.ps1
Step 2. Initialize a new Spice app with the spice init
command:
spice init spice_qs
A spicepod.yaml
file is created in the spice_qs
directory. Change to that directory:
cd spice_qs
Step 3. Start the Spice runtime:
spice run
Example output will be shown as follows:
Spice.ai runtime starting...
2024-08-05T13:02:40.247484Z INFO runtime::flight: Spice Runtime Flight listening on 127.0.0.1:50051
2024-08-05T13:02:40.247490Z INFO runtime::metrics_server: Spice Runtime Metrics listening on 127.0.0.1:9090
2024-08-05T13:02:40.247949Z INFO runtime: Initialized results cache; max size: 128.00 MiB, item ttl: 1s
2024-08-05T13:02:40.248611Z INFO runtime::http: Spice Runtime HTTP listening on 127.0.0.1:8090
2024-08-05T13:02:40.252356Z INFO runtime::opentelemetry: Spice Runtime OpenTelemetry listening on 127.0.0.1:50052
The runtime is now started and ready for queries.
Step 4. In a new terminal window, add the spiceai/quickstart
Spicepod. A Spicepod is a package of configuration defining datasets and ML models.
spice add spiceai/quickstart
The spicepod.yaml
file will be updated with the spiceai/quickstart
dependency.
version: v1beta1
kind: Spicepod
name: spice_qs
dependencies:
- spiceai/quickstart
The spiceai/quickstart
Spicepod will add a taxi_trips
data table to the runtime which is now available to query by SQL.
2024-08-05T13:04:56.742779Z INFO runtime: Dataset taxi_trips registered (s3://spiceai-demo-datasets/taxi_trips/2024/), acceleration (arrow, 10s refresh), results cache enabled.
2024-08-05T13:04:56.744062Z INFO runtime::accelerated_table::refresh_task: Loading data for dataset taxi_trips
2024-08-05T13:05:03.556169Z INFO runtime::accelerated_table::refresh_task: Loaded 2,964,624 rows (421.71 MiB) for dataset taxi_trips in 6s 812ms.
Step 5. Start the Spice SQL REPL:
spice sql
The SQL REPL inferface will be shown:
Welcome to the Spice.ai SQL REPL! Type 'help' for help.
show tables; -- list available tables
sql>
Enter show tables;
to display the available tables for query:
sql> show tables;
+---------------+--------------+---------------+------------+
| table_catalog | table_schema | table_name | table_type |
+---------------+--------------+---------------+------------+
| spice | public | taxi_trips | BASE TABLE |
| spice | runtime | query_history | BASE TABLE |
| spice | runtime | metrics | BASE TABLE |
+---------------+--------------+---------------+------------+
Time: 0.022671708 seconds. 3 rows.
Enter a query to display the longest taxi trips:
SELECT trip_distance, total_amount FROM taxi_trips ORDER BY trip_distance DESC LIMIT 10;
Output:
+---------------+--------------+
| trip_distance | total_amount |
+---------------+--------------+
| 312722.3 | 22.15 |
| 97793.92 | 36.31 |
| 82015.45 | 21.56 |
| 72975.97 | 20.04 |
| 71752.26 | 49.57 |
| 59282.45 | 33.52 |
| 59076.43 | 23.17 |
| 58298.51 | 18.63 |
| 51619.36 | 24.2 |
| 44018.64 | 52.43 |
+---------------+--------------+
Time: 0.045150667 seconds. 10 rows.
Using the Docker image locally:
docker pull spiceai/spiceai
In a Dockerfile:
from spiceai/spiceai:latest
Using Helm:
helm repo add spiceai https://helm.spiceai.org
helm install spiceai spiceai/spiceai
You can use any number of predefined datasets available from the Spice.ai Cloud Platform in the Spice runtime.
A list of publicly available datasets from Spice.ai can be found here: https://docs.spice.ai/building-blocks/datasets.
In order to access public datasets from Spice.ai, you will first need to create an account with Spice.ai by selecting the free tier membership.
Navigate to spice.ai and create a new account by clicking on Try for Free.
After creating an account, you will need to create an app in order to create to an API key.
You will now be able to access datasets from Spice.ai. For this demonstration, we will be using the spice.ai/eth.recent_blocks
dataset.
Step 1. Initialize a new project.
# Initialize a new Spice app
spice init spice_app
# Change to app directory
cd spice_app
Step 2. Log in and authenticate from the command line using the spice login
command. A pop up browser window will prompt you to authenticate:
spice login
Step 3. Start the runtime:
# Start the runtime
spice run
Step 4. Configure the dataset:
In a new terminal window, configure a new dataset using the spice dataset configure
command:
spice dataset configure
You will be prompted to enter a name. Enter a name that represents the contents of the dataset
dataset name: (spice_app) eth_recent_blocks
Enter the description of the dataset:
description: eth recent blocks
Enter the location of the dataset:
from: spice.ai/eth.recent_blocks
Select y
when prompted whether to accelerate the data:
Locally accelerate (y/n)? y
You should see the following output from your runtime terminal:
2024-08-05T13:09:08.342450Z INFO runtime: Dataset eth_recent_blocks registered (spice.ai/eth.recent_blocks), acceleration (arrow, 10s refresh), results cache enabled.
2024-08-05T13:09:08.343641Z INFO runtime::accelerated_table::refresh_task: Loading data for dataset eth_recent_blocks
2024-08-05T13:09:09.575822Z INFO runtime::accelerated_table::refresh_task: Loaded 146 rows (6.36 MiB) for dataset eth_recent_blocks in 1s 232ms.
Step 5. In a new terminal window, use the Spice SQL REPL to query the dataset
spice sql
SELECT number, size, gas_used from eth_recent_blocks LIMIT 10;
The output displays the results of the query along with the query execution time:
+----------+--------+----------+
| number | size | gas_used |
+----------+--------+----------+
| 20462425 | 32466 | 6705045 |
| 20462435 | 262114 | 29985196 |
| 20462427 | 138376 | 29989452 |
| 20462444 | 40541 | 9480363 |
| 20462431 | 78505 | 16994166 |
| 20462461 | 110372 | 21987571 |
| 20462441 | 51089 | 11136440 |
| 20462428 | 327660 | 29998593 |
| 20462429 | 133518 | 20159194 |
| 20462422 | 61461 | 13389415 |
+----------+--------+----------+
Time: 0.008562625 seconds. 10 rows.
You can experiment with the time it takes to generate queries when using non-accelerated datasets. You can change the acceleration setting from true
to false
in the datasets.yaml file.
Comprehensive documentation is available at docs.spiceai.org.
Spice.ai is designed to be extensible with extension points documented at EXTENSIBILITY.md. Build custom Data Connectors, Data Accelerators, Catalog Connectors, Secret Stores, Models, or Embeddings.
🚀 See the Roadmap to v1.0-stable for upcoming features.
We greatly appreciate and value your support! You can help Spice in a number of ways:
- Build an app with Spice.ai and send us feedback and suggestions at hey@spice.ai or on Discord, X, or LinkedIn.
- File an issue if you see something not quite working correctly.
- Join our team (We’re hiring!)
- Contribute code or documentation to the project (see CONTRIBUTING.md).
- Follow our blog at blog.spiceai.org
⭐️ star this repo! Thank you for your support! 🙏