Any AI agent that connects to the Realmint MCP can instantly look up risk scores, read issuer fine print, pull live prices, and trace metadata changes across 140+ tokenized real-world assets. No API key needed for the read-only catalog tools, no installation, just a single URL. This post covers what the server exposes, how to connect it in under a minute, and what both agents and people can do with it. The quick-start page lives at realmint.io/developers/mcp.
What is MCP?
MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude call external tools natively, as structured function calls rather than scraped text. When you connect the Realmint MCP to Claude, it gains new tools it can invoke directly, without copy-pasting, without browser use, and without any prompt engineering. The agent calls the tool and gets clean, structured data back.
The Realmint MCP server is hosted at mcp.realmint.io/mcp. Its read-only catalog tools are public and require no account or API key; a separate set of OAuth-gated execution tools lets an authenticated agent buy, sell, and withdraw.
How to connect it
Go to Settings → Connectors → Add custom connector. Paste https://mcp.realmint.io/mcp and save. The read-only tools appear immediately in any new conversation; the execution tools prompt for a one-time OAuth sign-in on first use.
Run claude mcp add realmint https://mcp.realmint.io/mcp in your terminal. The tools are available in every Claude Code session from that point on.
Any client that supports Streamable HTTP transport can connect. Point it at https://mcp.realmint.io/mcp. No auth required for the read-only catalog tools.
What your agent can do
Once connected, the agent has access to a full RWA research stack. Here are the most common workflows:
Research any tokenized asset before you buy
Ask the agent to look up an asset and it can pull its complete due diligence profile in one call: who issued it, what jurisdiction it operates under, whether the issuer can freeze or blacklist wallets, what the redemption terms are, and how it scores across 6 independent risk dimensions. This is information that normally requires reading a 40-page legal document.
Check whether an asset can actually be traded
The agent can query which venues and chains currently support a given asset before quoting any position. If something is delisted or the routing provider is unhealthy, it gets a structured signal to stop, not a failed transaction.
Catch when issuers quietly change the rules
Every metadata update to every asset is stored as a versioned snapshot. An agent can diff two versions and surface changes to control parameters, redemption terms, or governing law. This is exactly the kind of thing that never shows up in the price chart but matters a lot if you're holding.
Compare assets across risk and price
Pull live market data and risk scores for a whole category in two calls, then sort and filter by composite score, liquidity, market cap, or 24h change. The structured output means the agent gets clean JSON it can reason over directly, with no text parsing needed.
The read-only research tools
realmint_list_assets
Returns the full catalog. Each entry includes asset ID, type, chain, contract address, issuer details, on-chain control capabilities (freeze, mint, burn, blacklist), transferability restrictions, redemption mechanics, and links to legal documents. Check metadata.lifecycle on each asset: a non-null value means the token is discontinued or migrated.
realmint_get_asset
Returns the full schema for a single asset by its lowercased ticker ID (e.g. xaut, paxg, tslax). Same structure as a catalog entry, useful for targeted lookups. If metadata.lifecycle is non-null, the asset is inactive and its risk score is hard-capped at 49.
realmint_get_score
Returns the Realmint Score: individual scores for all 6 risk dimensions plus a weighted composite (0–100, higher is better). Liquidity carries the most weight (25%), followed by social presence (22%), exit rights (18%), backing quality (14%), enforceability (14%), and issuer control risk (7%).
Score thresholds
realmint_get_asset_history
Returns a versioned changelog of every metadata update to an asset, with a full schema snapshot at each revision. Use this for ongoing monitoring: if freeze capabilities or redemption restrictions appear in a revision that didn't have them before, that's a material change worth flagging.
realmint_get_route_support
Returns whether an asset is routable via Realmint and lists all available venues with their chain, type (DEX AMM, RFQ, issuer primary, lending), provider ID, and health status. Call this before attempting any trade. If routing_enabled is false, the asset can be researched but not traded through Realmint.
realmint_list_market
Returns live market data (spot price, 24h change, volume, market cap, and venue count) for one or many assets. Pass a comma-separated list of IDs to filter, e.g. xaut,paxg, or leave it empty for the full catalog.
realmint_get_price_history
Returns time-series price data with a configurable lookback of 1–730 days and granularity of 5-minute, 1-hour, or 1-day buckets. The server auto-selects granularity if omitted: 5m for ≤2 days, 1h for ≤30 days, and 1d beyond that. Useful for trend analysis, volatility modelling, or backtesting.
What assets are covered
The Realmint catalog currently covers 140+ assets across two main categories, with more on the way:
- Tokenized equities. 140+ tokenized US stocks and ETFs, including TSLAx, AAPLx, NVDAx, METAx, MSFTx, GOOGLx, AMZNx, PLTRx, AMDx, CRMx, JPMx, HOODx, COINx, MSTRx, QQQx, and more. Issued primarily by Backed (xStocks), native to Solana and accessible from USDC on Injective via cross-chain routing.
- Tokenized commodities. Physical commodities including gold, silver, uranium, copper, and more, each backed 1:1 by real-world holdings and indexed with full issuer and custody details.
Tokenized collectibles are live too: over 3,000 Collector Crypt trading cards are indexed and tradeable. Additional categories, including tokenized yield instruments, are in active development and will be added to the catalog as they launch.
Every asset is indexed with a complete schema covering issuer identity, on-chain control powers, transferability restrictions, redemption terms, regulatory framework, and linked legal documents. The catalog is updated continuously.
About the Realmint Score
The score returned by realmint_get_score is produced by an automated model that reads publicly available data: smart contract parameters, audit reports, legal documents, and issuer communications. It scores each asset across 6 dimensions. Scores are recomputed as data changes, so an agent polling periodically gets a live signal rather than a registration-time snapshot. The policy_version field in every response lets agents distinguish a score change caused by new asset data versus a model update.
For the full scoring methodology, see How to Evaluate Any Tokenized Asset: 6 Risk Dimensions Explained.
The Realmint Score is under active development. Weights, penalties, and data sources change as the model improves. Treat scores as a structured research input, not a guarantee. Always verify material facts with primary sources before making any investment decision.
Read-only catalog, OAuth-gated execution
The catalog tools are read-only and need no authentication — connecting an agent to just those carries no execution risk, since the worst they can do is read data. On top of them, the server exposes OAuth-gated execution tools (deposit, buy, sell, withdraw): after a one-time OAuth sign-in and wallet delegation, Realmint signs on the agent's behalf server-side, so no private key ever touches the agent. The execution tools only appear once an agent authenticates, so a read-only integration stays read-only.
The full quick-start guide, with setup instructions for every major MCP client, is at realmint.io/developers/mcp.
Browse all indexed RWA assets, their risk scores, and live market data, the same data the MCP exposes, in the Realmint UI.
Explore the Realmint catalog