Platform Coverage

Embed's V3 platform covers three domains — Social, Coins, and Prediction Markets — across multiple protocols and chains. This page shows what's available today and what's coming next.

Coverage Overview

Domain	Protocol	Search Indices	Features	Ranking	Hydration
Social	Farcaster	`farcaster-users`, `farcaster-items`	Social Affinity, Content Sims, AI Labels	Session-Based, Online Learning	—
	Zora	posts, users, trades	Social Affinity, AI Labels, Trader Affinity, Content Sims	Session-Based, Online Learning	—
Coins	Ethereum	coins, users	Social Affinity, Trader Affinity, Content Sims	Session-Based, Online Learning	`interesting_trades`, `convergence_trades`, `interesting_price_change`, `user_balances`
	Solana	coins, users	Social Affinity, Trader Affinity, Content Sims	Session-Based, Online Learning	`interesting_trades`, `convergence_trades`, `interesting_price_change`, `user_balances`
	Base	coins, users	Social Affinity, Trader Affinity, Content Sims	Session-Based, Online Learning	`interesting_trades`, `convergence_trades`, `interesting_price_change`, `user_balances`, `user_recent_activity`
Prediction Markets	Polymarket	`polymarket-wallets`, `polymarket-items`	Social Affinity, Trader Affinity, Content Sims, AI Labels	Session-Based, Online Learning	`interesting_bets`, `interesting_price_change`, `user_recent_activity`, `convergence_bets`
	Kalshi	`kalshi-wallets`, `kalshi-items`	Social Affinity, Trader Affinity, Content Sims, AI Labels	Session-Based, Online Learning	`interesting_bets`, `interesting_price_change`, `user_recent_activity`, `convergence_bets`
	Perps / Stocks / Commodities	TBD	—	—	—

Domains

Social

Farcaster — decentralized social protocol. Two search indices:

farcaster-users — user profiles
farcaster-items — posts (casts)

Features include Social Affinity (who the user follows and interacts with), Content Similarity (embedding-based matching), and AI Labels (topic, sentiment, emotion, moderation). Ranking models are backtested across both session-based and online learning approaches.

Zora — creator economy platform. Three data types (posts, users, trades) are indexed. Features cover Social Affinity, AI Labels, Trader Affinity, and Content Similarity.

Coins

Token trading activity across three chains, all following the same architecture:

Chain	Unique capabilities
Ethereum	Highest liquidity, most established DeFi
Solana	Highest memecoin activity
Base	Growing DEX activity, includes `user_recent_activity` hydration

Each chain provides:

Search indices — coins and users
Features — Social Affinity, Trader Affinity, Content Sims
Ranking — Session-Based and Online Learning
Hydration — interesting_trades, convergence_trades, interesting_price_change, user_balances (real-time enrichment from DynamoDB)

The multi-chain token data is also available via the token-items search index, which aggregates tokens across all three chains. See Supported Indices for the full field reference.

Prediction Markets

Polymarket — the most mature prediction market integration. Full feature coverage including AI Labels, all affinity features, and both ranking approaches. Hydration provides interesting_bets, convergence_bets, interesting_price_change, and user_recent_activity.

Kalshi — prediction markets on Solana via DFlow contracts. Same architecture as Polymarket with equivalent feature and ranking coverage. Hydration provides the same signals.

Perps / Stocks / Commodities — planned, not yet available.

Service Layers Explained

Search (Elasticsearch)

The search layer retrieves candidates from indexed data. Each protocol has one or more search indices containing items and user profiles. Search supports structured filtering, relevance boosting, and semantic similarity.

See [Search Guide](../Building Algorithms/search) and Supported Indices for the full field reference per index.

Features (Elasticsearch)

The features layer enriches candidates with ML-computed signals by comparing items against a user profile:

Feature	What it measures
Social Affinity	How connected the user is to the item's creator via social graph
Trader Affinity	How similar the user's trading behavior is to other traders on this item
Content Sims	Semantic similarity between the item and the user's interaction history
AI Labels	Topic, sentiment, emotion, moderation, and Web3 classifications

See [Features Guide](../Building Algorithms/algo-features) for API details.

Ranking (VW / Vowpal Wabbit)

The ranking layer applies trained ML models to produce personalized orderings:

Approach	Description
Session-Based	Models trained on per-session interaction patterns
Online Learning	Models that update continuously from real-time user feedback

See [Scoring & Ranking Guide](../Building Algorithms/algo-scoring-ranking) for the sort, linear, and mix methods.

Hydration (DynamoDB)

The hydration layer provides real-time enrichment signals from DynamoDB, primarily for Coins and Prediction Markets:

Signal	Domain	Description
`interesting_trades`	Coins	Notable recent trades by smart money wallets
`convergence_trades`	Coins	Multiple top wallets trading the same token
`interesting_price_change`	Coins, Prediction Markets	Significant price movements
`user_balances`	Coins	Current token holdings for a wallet
`user_recent_activity`	Coins (Base), Prediction Markets	Recent trading activity for a user
`interesting_bets`	Prediction Markets	Notable bets by top-ranked traders
`convergence_bets`	Prediction Markets	Multiple top traders betting on the same market

These signals power the Smart Money Signals API and can be used to build real-time notification feeds.

What's Next

Supported Indices → Full field reference for every search index
[Search →](../Building Algorithms/search) How to query indices with filters, boosting, and semantic search
[Features →](../Building Algorithms/algo-features) ML enrichment and personalization signals
Smart Money Signals → Real-time smart money intelligence powered by hydration signals