Blocktronics launches V2 with cost basis tracking and AI agent tools

Blocktronics launched V2 of its onchain token intelligence platform on October 28, 2026, introducing cost basis for every wallet, smart money weighted by track record, and an API and Model Context Protocol server for AI agents.

The company announced the release in a post on X, describing V2 as a rebuilt system for onchain data distribution. Users top up or perform swaps to earn credits, and the API and MCP plug-ins are now live, according to Blocktronics.

The platform covers 233 tokens on Base, Ethereum, and Robinhood Chain.

What a token page shows

The centerpiece is a comparison of holder value against cost basis. Holder value reflects the current worth of tokens held, while cost basis, also called realized cap, represents the total amount paid by those holders. When holder value exceeds cost basis, holders are in unrealized profit; when it falls below, they are in unrealized loss. This relationship is expressed as MVRV, or holder value divided by cost basis.

Blocktronics states that cost basis is tied to actual purchase prices, making it less volatile than token price and indicating the holder base’s position. It updates whenever tokens change hands at a new price.

Each token page also includes DEX netflow, smart money netflow, and a Blocktronics grade. The grade, ranging from A to E, assesses structural fragility based on six pillars: depth, concentration, holders, flow, economics, and launch. The company clarifies that the grade is not a price prediction.

V2 adds cost basis for individual wallets and allows users to filter out exchange, team, or escrow wallets. Exchange and infrastructure wallets are excluded from holder metrics, as they represent multiple users rather than single holders. Capital totals still include exchange balances, as these tokens belong to investors. Liquidity pools, bridges, and protocol contracts are excluded from both metrics.

Smart money, weighted by record

Blocktronics defines smart money as a group of wallets identified by its backend as proven performers, selected based on past behavior rather than future actions. The company’s documentation states that size alone does not indicate skill, and large wallets do not automatically qualify.

V2 weights this cohort by track record, though the selection criteria remain private to prevent gaming, according to Blocktronics. The public view shows the cohort’s flows and holdings for each token but does not list the wallets.

An API and MCP server for agents

The release includes an MCP server with over 40 tools for AI agents. Agents can search tokens, retrieve overviews, pull 60-day series, compare up to 12 tokens, screen coverage, read signals, render chart cards, and build dashboards and hubs. Most functions, including reads, charts, dashboards, and hubs, require no account.

Blocktronics states the server is compatible with Claude, ChatGPT, Cursor, VS Code, Codex, and other clients.

Agents can obtain an API key without human intervention by signing a message with their wallet. They then top up credits in USDC from their own wallet. According to the documentation, refusals, errors, and quotes are free, and each response includes the cost. A token series costs 2 credits per token, while compare and screen calls are priced by size.

The API follows the same pricing model. A catalog endpoint lists all daily metrics, token page sections, and tools with their inputs and prices, and does not require a key. Agents can query up to five tokens in one request or ingest a token not yet indexed by Blocktronics, building it from its first transfer.

Other tools in the release

Blocktronics also introduced a Bundle Detector, which identifies wallets that bought a token at launch together and tracks their remaining holdings against liquidity. A bundle scan costs 5 credits and remains available for 24 hours. A new-token scan covers any token not yet indexed by Blocktronics, up to 250 credits. A daily market brief and a Telegram bot, @BlocktronicsBot, deliver token cards and market updates in chat.

For AI agents researching or trading tokens, the key change is that holder data is now accessible and payable programmatically. V2 is designed for both human and machine use, with pricing structured so agents can determine costs before committing to a query.