--- name: yddkctoq.nx.link description: This host provides two complementary blockchain-focused capabilities: a natural-language-to-EVM-transaction planner powered by thirdweb's t0 LLM, and a token metadata lookup service that returns symbol, decimals, USD price, and related data across multiple chains. Together they support the planning and parameterization of on-chain DeFi operations without executing transactions directly. host: yddkctoq.nx.link --- # yddkctoq.nx.link yddkctoq.nx.link serves agents that need to translate human-readable blockchain intents into structured EVM execution plans and resolve token metadata needed to correctly parameterize those plans. It is not a general-purpose LLM host, a trade executor, or a real-time price feed. Its two skills are narrowly scoped to the pre-execution phase of DeFi workflows on EVM chains. ## When to use this host Use this host when an agent needs to (1) look up token metadata such as decimals or USD price across EVM chains, or (2) convert a natural-language DeFi intent into a structured EVM transaction plan with calldata. Do not use it for general text generation — use a standard OpenAI-compatible LLM host instead. Do not use it for live price streaming, order book data, or swap routing — dedicated DEX aggregator or price-feed hosts are more appropriate. Critically, this host does not execute or sign transactions; a separate on-chain execution layer is required after consuming its outputs. ## Capabilities ### Token Metadata Resolution Resolves token identity and pricing data — symbol, name, decimals, USD price, and multi-currency conversions — for tokens identified by chain and address. Provides the ground-truth token parameters needed before constructing any transaction. - **`fetch-token-metadata`** — Returns token metadata including symbol, name, decimals, USD price, and multi-currency price conversions for tokens across multiple chains. ### Blockchain Transaction Planning Converts natural-language DeFi or wallet goals into structured EVM transaction plans, including contract addresses, function calldata, slippage logic, and multi-step sequencing. Produces a plan ready for downstream signing and submission. - **`chat-thirdweb-t0-blockchain`** — Sends an OpenAI-compatible chat completion request to thirdweb's t0 LLM, which returns structured blockchain transaction plans, on-chain read sequences, and EVM calldata for natural-language DeFi and wallet tasks. ## Workflows ### Token-Aware Transaction Planning *Use when an agent needs to build an EVM transaction plan for a DeFi operation and must first confirm token decimals, current USD price, or symbol before passing accurate parameters to the planner.* 1. **`fetch-token-metadata`** — Resolve the token's decimals, symbol, and current USD price for each token involved in the intended operation, ensuring accurate parameterization. 2. **`chat-thirdweb-t0-blockchain`** — Pass the resolved token metadata alongside the natural-language goal so the t0 LLM can produce a correctly scaled, decimal-aware EVM transaction plan. ## Skill reference ### `chat-thirdweb-t0-blockchain` **thirdweb t0 Blockchain Chat** — Sends an OpenAI-compatible chat completion request to thirdweb's t0 LLM, which returns structured blockchain transaction plans, on-chain read sequences, and EVM calldata for natural-language DeFi and wallet tasks. *Use when:* Use when an agent needs to convert a natural-language blockchain goal into an executable EVM transaction plan, including contract addresses, function calldata, token decimal handling, slippage logic, and safety checks — especially for multi-step DeFi flows on EVM chains. *Not for:* Do not use for general-purpose text generation or non-blockchain tasks; use a standard OpenAI-compatible LLM instead. Not suitable for real-time price feeds or live on-chain execution — this returns a plan, not a signed transaction. **Inputs:** - `model` (string, required) — Model identifier for the thirdweb t0 LLM. - `messages` (array, required) — Array of OpenAI-compatible chat message objects, each with 'role' (system, user, assistant) and 'content' (string). Include a system prompt describing the task context and a user message with the blockchain goal. **Returns:** Returns a JSON-encoded message with findings, on_chain_reads, computed_values (e.g. amountIn in base units), a transaction_plan with exactInputSingle calldata and gas estimates, safety_checks, and error_handling for the requested EVM flow. **Example:** `{"model":"thirdweb-t0-chat","messages":[{"role":"system","content":"You are an OpenAI-compatible blockchain assistant. Convert natural-language goals into executable thirdweb/EVM steps."},{"role":"user","content":"On Polygon (chainId 137), swap 500 USDC (0x2791Bca1f2de4661ED88A30C99A7a9449Aa84174) to WETH (0x7ceB23fD6bC0adD59E62ac25578270cFf1b9f619) using Uniswap V3 with 0.5% slippage. Return a full transaction plan."}]}` --- ### `fetch-token-metadata` **TokenFinder** — Returns token metadata including symbol, name, decimals, USD price, and multi-currency price conversions for tokens across multiple chains. *Use when:* Use when an agent needs token metadata such as price in USD or other fiat currencies, decimals, market cap, or 24h volume for one or more tokens identified by chain and address. *Not for:* Do not use for real-time streaming price feeds or order book data; this is a single-shot metadata lookup. Not suitable for swap routing or executing trades. **Inputs:** - `query` (string) — Search query to filter tokens by address, name, or symbol. Exact parameter name not confirmed by schema; inferred from description. **Returns:** Returns a result object containing a tokens array, each with chainId, address, symbol, name, decimals, priceUsd, optional volume24hUsd and marketCapUsd, and a prices map with fiat conversions across ~20 currencies. **Example:** `GET https://yddkctoq.nx.link/v1/tokens` ---