📡 PreStocks Pulse

The open data layer for tokenized pre-IPO stocks on Solana. Premium radar · PRE8 index · creation/redemption flows · split-aware executable prices · an MCP server so AI agents can research and trade PreStocks.

• Live dashboard: https://aralroca.github.io/prestocks-pulse/
• Open dataset (hourly, JSON): https://aralroca.github.io/prestocks-pulse/data/history.json
• MCP server: npx -y github:aralroca/prestocks-pulse (Node ≥ 23.6 or Bun)

Built for the Stocklana hackathon, Best Use of PreStocks bounty.

The problem

PreStocks tokens (OPENAI, ANTHROPIC, SPACEX, ANDURIL, …) trade 24/7 on Solana, but a holder has no easy way to answer the questions a Bloomberg terminal answers for public stocks:

1. Am I paying a premium? The token price and the SPV mark price (fair value) diverge — sometimes by double digits.
2. What price will I actually get? The issuer API price is an indicative number; the executable price lives on Jupiter and depends on liquidity.
3. Is money flowing in or out? There is no history API, so nobody can see supply changes (creations/redemptions), the pre-IPO equivalent of ETF fund flows.
4. Can my AI agent use any of this? Nothing exposes PreStocks to Claude, Cursor or any MCP client.

What Pulse does

Use it from your AI agent

Claude Desktop / Claude Code / Cursor (mcp.json):

{
  "mcpServers": {
    "prestocks-pulse": { "command": "npx", "args": ["-y", "github:aralroca/prestocks-pulse"] }
  }
}

Claude Code one-liner:

claude mcp add prestocks-pulse -- npx -y github:aralroca/prestocks-pulse

Then ask things like:

“Which PreStocks trade at a discount to their mark price right now?” “Quote 500 USDC of ANTHROPIC and tell me the executable premium.” “Build a $2,000 PRE8 basket.” “Did OPENAI see net creations or redemptions in the last 24 hours?”

Tools

The server reads the local dataset when run inside the repo and the published dataset otherwise, so it works from any directory.

Dataset schema

docs/data/history.json

{
  "version": 1,
  "source": "https://prestocks.com/api/prestocks",
  "snapshots": [
    {
      "t": "2026-09-17T19:17:51.560Z",
      "tokens": {
        "OPENAI": { "p": 1045.55, "m": 965.78, "s": 2826.5, "mv": 1541000000000 }
        // p = token price, m = mark price, s = supply, mv = mark valuation (USD)
      }
    }
  ]
}

docs/data/prestocks.json holds the latest full catalogue (names, descriptions, logos, mints) plus the ScaledUiAmount multipliers read from Solana.

Architecture

prestocks.com/api ──hourly──▶ GitHub Action ──▶ docs/data/*.json ──▶ GitHub Pages
                                                      │                    │
Jupiter (price v3 + quote) ◀─────── browser ◀─────────┘         dashboard (docs/)
Solana RPC (Token-2022 ScaledUiAmount) ◀── MCP server (src/mcp) ◀── Claude / Cursor

• src/metrics.ts — pure analytics (premium, index, flows, basket, divergence), shared by the dashboard and the MCP server.
• src/jupiter.ts — Jupiter price + quote client, multiplier-aware.
• src/solana.ts — reads scaledUiAmountConfig from the mints.
• src/history.ts — dataset read/append (local file or published URL).
• src/mcp/ — MCP server (@modelcontextprotocol/sdk, stdio).
• src/web/ — dashboard, bundled with bun build into docs/app.js.

Development

bun install
bun test              # unit + MCP end-to-end (in-memory transport)
bun run snapshot      # append one snapshot to docs/data
bun run build         # bundle the dashboard
bun run dev           # build + serve docs/ locally
bun run mcp           # run the MCP server on stdio

Why Solana

PreStocks only exist on Solana: Token-2022 mints with transfer fees, permanent delegate and ScaledUiAmount for corporate actions, traded on Jupiter-routed venues (Meteora DLMM, Manifest, Whirlpool). Pulse reads those primitives directly, which is what makes split-aware executable prices possible.

Disclaimer

Not investment advice. PreStocks provide economic exposure only and are not available to U.S. persons; read prestocks.com before trading. Open-source components: @modelcontextprotocol/sdk, zod.

License

MIT