Index Network x Ceramic Network: Decentralized AI with Composable Data

An introduction to Index Network

Index helps connect people with meaningful information, helping to solve problems that cause current web discovery tools to fall short. Instead of searching the entire internet, reviewing information without proper context, or relying on algorithms heavily influenced by monetization and optimization, Index empowers users to create custom, highly personalized search engines, called indexes.

Indexes are composable, searchable, and queryable with natural langauge, focused on specific data chosen by the user that creates them, such as documents, products, and other data types. They allow for direct discovery across peers, eliminating the need for intermediaries to help find information, products, and like-minded people. And thanks to AI, which assists with predictive search and semantic comprehension, search results are powerful and relevant.

Going deeper

Index enables interactions for new semantic protocols in a web3 verifiable data ecosystem. These include identity, reputation, intent, and even intuition. By creating a contextual environment for these protocols to interact with natural language and be discovered together, Index allows autonomous agents to interact with the network in a reactive, event-driven way. This, in turn, unlocks coordination and discovery across agents.

Here’s an example: below, Index captures knowledge, context, identity, and intent, all in one sentence. Responses like these can be used for ongoing information discovery in a novel, user-centric way.

How Ceramic powers Index

Index is built on Ceramic's decentralized graph database, ComposeDB, leveraging its capabilities to promote decentralization and enhance interoperable ownership and discovery.

Here’s how Index uses Ceramic’s infrastructure to power decentralized AI:

Decentralized Semantic Index: a composable vector database on ComposeDB, enabling semantic interoperability

Index enables discovery through a composable vector database, called a “decentralized semantic index.” Index stores vector embeddings on the same graph with the indexed data, Ceramic’s decentralized knowledge graph solution: ComposeDB. This approach allows modularity for a highly relational composition of use cases.

By combining Index's AI-based discovery functions with Ceramic's data ownership and decentralized network, Index becomes a decentralized layer for information discovery across peers. This ensures both data integrity and privacy while distributing trust among various nodes. Using Ceramic, Index Network avoids the privacy risks often found in centralized systems, and allows user-centric, composable information discovery across peers.

Imagine, for example, a chat setup that generates responses from both a community index and a private index. This type of connected discovery experience can deliver responses that are simultaneously personalized and trusted.

Composable Discovery Protocol: Real-time, event-driven coordination for autonomous agents, powered by Ceramic streams

Let’s talk about the algorithm part. Here’s how it works: Index’s discovery protocol leverages Ceramic to orchestrate a network of algorithms, functioning as autonomous agents. These agents operate with a range of diverse economic models, communicating and interacting based on the knowledge they gather and share. Ceramic’s infrastructure supports these capabilities by enabling subscriptions to dynamic data streams. This is crucial for AI agents that need to react to changes in user contexts or specific conditions in real-time.

Here’s another example: using Ceramic's event-driven architecture, an AI agent might subscribe to updates with natural language for a search like "new effects of quantum material science on computing technologies." This would allow it to execute tasks or provide information as soon as relevant data becomes available.

In this way, Ceramic’s capabilities not only help enhance the responsiveness of AI agents, but also ensure that they operate in a privacy-preserving and user-centric manner. This allows Index Network to facilitate the development of a reactive, real-time AI environment, ready to adapt to the evolving demands of future economies.

Why is this only possible with a composable data network? And why Ceramic?

Aside from the crucial technological capabilities provided by Ceramic, we’ve found that the support of the Ceramic team has been vital. We’ve had a collaborative relationship from the very beginning, which has allowed us to test, learn and iterate at speed as we build an entirely new paradigm for information discovery on the internet.

Additionally, building on Ceramic allows us to operate with decentralized data ownership, leveraging DIDs from the get-go in a user-centric way.

The Index ecosystem

We acknowledge the importance of building collaboratively if we hope to realize our vision. The new internet will not be a product of any single company, but rather the result of an open ecosystem of multiple entities, protocols, and algorithms.

Today, Index is developed using Ceramic’s ComposeDB, which integrates indexed data as part of an open, adaptable data ecosystem. All data is stored on IPFS, empowering users to create, own, and manage their data independently. We partnered with Lit Protocol to ensure decentralized access control with blockchain-backed privacy and also partnered with Fluence Network to enable search and discovery on a decentralized computation network, ensuring that user discoveries are kept private thanks to a composable privacy approach.

Combining these essential technologies into a single framework creates the foundational infrastructure needed for healthy and more relevant web discovery. But wait, there’s more. Index partnered with identity, trust, and reputation protocols, which form the cornerstone for navigating data within Index’s autonomous structure. On the identity front, Disco provides discovery for all the contexts users have; with Intuition Systems, users can build and explore trust; Verax allows users to sense reputations, and with Olas Network, to enable the network to services and offerings provided by autonomous agents. For Index, all these models are part of a user’s context, and taken together, they form the fundamental building blocks for creating a truly personalized internet experience.

Here’s a preview:

https://youtu.be/YBJ6IZhundw

Future plans

Index is now working to facilitate search and discovery within private data. We’re handling this by storing vector embeddings on ComposeDB, and leveraging decentralized computation through the Fluence Network. This approach aims to create a unified discovery platform that can work with varying privacy levels.

Index is also exploring cross-model embedding spaces to foster interoperability among AI models. This initiative seeks to enable collaboration and communication between different AI models, allowing for seamless transitions across different models.

If you’re interested in decentralized information discovery and what we’re building at Index Network, we’d love to hear from you.