The Spice AI Blog

Data is the fuel for AI-driven apps that learn and adapt. We’re building an application platform that combines the open source spiceuseai.us decision-engine with data enabling developers to build the next-generation of intelligent applications.

Spice.xyz June Update

3 min readJun 21, 2022

--

Press enter or click to view image in full size
Spice.xyz June 2022 Update

Since announcing Spice.xyz in April, we’ve made a ton of updates and are excited to announce:

These updates make it even easier for developers, data-scientists, and AI/ML engineers to leverage bulk, web3 data in their applications, notebooks, and pipelines.

For example, now, with the Python SDK, it’s takes only three lines of Python code, including your query, to get web3 data into pandas and NumPy formats.

The new SDKs stream data over high-performance Apache Arrow Flight connections enabling you to query and use millions of rows of web3 data in your applications, notebooks, and ML pipelines in seconds.

Python SDK

The Spice Python SDK is the easiest way to use and query Spice from Python.

With the SDK, the vast ecosystem of Python tools and libraries for data-science and ML are now available to you, like pandas, NumPy, PyTorch, TensorFlow, scikit-learn, and more.

Here’s an example of querying NFT Airdrops on Kaggle.

Javascript/Typescript SDK for Node.js

The new Node.js SDK provides a similarly easy experience.

defly.ai, the Ethereum gas fee optimizer we built last year, uses a Node.js backend. We wished we had the Node.js SDK when we created it, so we’re excited to migrate over to it ourselves.

Ethereum ERC-1155 Multi-Token Standard support

Support for ERC-1155 tokens has been a highly-requested feature. The ERC-1155 standard includes fungible and non-fungible tokens thus ERC-1155 is a requirement for use-cases from crypto tax accounting to NFT services.

Spice supports the full ERC-1155 standard, including tracking both fungible and non-fungible (NFT) tokens within each ERC-1155 contract.

New token standard datasets are also available such as eth.token_transfers_erc1155 and existing datasets like eth.tokens include ERC-1155 tokens.

Going forward we plan to enhance support for indexing ERC-721 and ERC-1155 metadata, even if it is stored off-chain, such as on IFPS or cloud storage.

Enhanced, automated token detection

ERC-1155 support was enabled through a complete rebuild of our automated token detection system, so ERC-20 and ERC-721 detection was also improved. The new systems works in realtime, detects more contracts and tokens, and overcomes traditional detection limitations such as proxy contracts.

Potential ERC-20, ERC-721, and ERC-1155 contracts are detected in real-time and with Spice providing a standards compliance score. This enables developers to be notified of conformance to each standard even as contracts are deployed or as tokens are minted.

Improvements to scale, reliability, and performance

Since the Spice.xyz preview launched in April, we’ve made significant scale, reliability, and performance improvements to support workloads, and ML pipelines in production. The Arrow Flight APIs easily handle millions of rows of data and we’ve enabled endpoints that can handle up to 1,000 RPS for pilot customers in our design partner program.

Because we’re designing Spice.xyz for real-time applications and ML, we’re constantly working to improve performance and scale.

Thanks for reading the first Spice.xyz Preview update — you can read detailed release notes at docs.spice.xyz. Being the first update, this is just the start. We’re growing the team, supporting more chains, like Solana, and working to make it even easier to leverage web3 data and AI in your application.

Learn More and Contribute

Even for advanced developers, building intelligent apps that leverage AI is still way too hard. Our mission is to make this as easy as creating a modern web page.

If that excites you, join us!

Try out Spice.xyz, get in touch on Discord, or reach out on Twitter.

Luke

--

--

The Spice AI Blog
The Spice AI Blog

Published in The Spice AI Blog

Data is the fuel for AI-driven apps that learn and adapt. We’re building an application platform that combines the open source spiceuseai.us decision-engine with data enabling developers to build the next-generation of intelligent applications.

Luke Kim
Luke Kim

Written by Luke Kim

Founder and CEO of Spice AI. Building for the next generation of intelligent web2 and web3 apps.

No responses yet