Sui View — Transaction Translator
Human-readable explanations for Sui blockchain transactions. Paste a transaction digest and get a clear, Markdown-formatted summary of what happened, which tokens were involved, object changes, and gas usage.
Architecture
Sui View is a web application that translates complex Sui blockchain transactions into easy-to-understand explanations. The system works in several stages to collect, enrich, and explain transaction data.
How It Works
When you submit a transaction digest, the application follows this process:
- Fetch Raw Transaction Data: Retrieves the complete transaction details from the Sui blockchain, including all balance changes, object modifications, and events.
 - Identify Tokens: Scans the transaction to find all cryptocurrency tokens involved, extracting their types and addresses.
 - Enrich with Metadata: Fetches human-readable information for each token (names, symbols, decimals) and transaction metadata (sender names, protocol information).
 - Convert to Human-Readable Format: Transforms raw numbers into readable amounts using proper decimal places and formatting.
 - Generate Explanation: Uses AI to analyze the enriched transaction data and create a natural language explanation of what happened.
 - Display Results: Presents the explanation in a clean, formatted Markdown display.
 
Key Components
The application consists of three main components:
- User Interface: A simple web interface where users can paste transaction digests or URLs. The interface handles input validation and displays the final explanation.
 - Data Enrichment Layer: Processes raw blockchain data by identifying tokens, fetching metadata, and converting technical values into human-friendly formats. This includes converting raw balance changes into readable amounts with proper token names and symbols.
 - Explanation Engine: Uses AI to analyze the enriched transaction data and generate natural language explanations. The engine is trained with Sui-specific context to provide accurate and relevant explanations.
 
Data Sources
Sui View relies on two primary data sources to provide comprehensive transaction explanations:
Blockberry Sui API
Blockberry provides on-chain data services for the Sui blockchain. Sui View uses Blockberry to:
- Fetch raw transaction data from the blockchain
 - Retrieve transaction metadata (sender names, protocol information)
 - Obtain coin metadata (token names, symbols, decimal places)
 
OpenAI
OpenAI's language model powers the explanation generation. The system uses GPT-4o to analyze enriched transaction data and create natural, human-readable explanations. The AI is provided with Sui-specific context to ensure accurate interpretation of transaction types and patterns.
Usage Guide
- Navigate to the home page and locate the transaction input field.
 - Paste a Sui transaction digest or a full transaction URL from Sui explorers.
 - Click the "Translate" button to begin processing.
 - Wait while the system fetches and analyzes the transaction data. You may see status messages indicating the progress.
 - Read the generated explanation, which will appear below the input field. The explanation includes details about tokens involved, balance changes, object modifications, and gas usage.
 
Tip: You can paste either a transaction digest (the alphanumeric hash) or a full URL from Sui Vision or other Sui explorers. The system will automatically extract the digest from URLs.
What You'll See
The explanation provides a comprehensive overview of the transaction:
- Transaction Summary: A high-level description of what the transaction accomplished
 - Token Movements: Details about which tokens were sent, received, or involved, with readable amounts and token names
 - Balance Changes: How account balances changed as a result of the transaction
 - Object Changes: Information about any Sui objects that were created, modified, or deleted
 - Gas Information: Details about the gas fees paid for the transaction
 - Events: Any events emitted by the transaction, which can provide additional context about what happened