Luxxon — TypeScript SDK + MCP server
Programmable vision for AI agents. Request live video or singleframes from real-world locations, settled on-chain in USDC.
This repo contains:
| Package | Purpose |
|---|---|
packages/sdk |
TypeScript HTTP client for the Luxxon API. Pure fetch, no native deps. |
packages/mcp |
MCP server exposing Luxxon as agent-callable tools (stdio transport, works with Claude Desktop / Cursor / any MCP client). |
examples/ |
Runnable agent examples — copy, paste, run. |
API: docs.luxxon.dev — https://api.luxxon.dev/api/v1.
Why an MCP server
Agents have fetch(), not <video> tags. The MCP server turns Luxxoninto a set of tools any MCP-aware agent can call directly:
get_frame(sessionId)→ JPEG bytes ready for a vision modelget_session(sessionId)→ current state of a live sessioncancel_session(sessionId)→ early termination
Drop the server into your Claude Desktop config and your agent cansee the world.
Quickstart — TypeScript SDK
import { Luxxon } from "@luxxon/sdk";
const lx = new Luxxon({ apiKey: process.env.LUXXON_API_KEY });
// Read the latest frame from a LIVE session (returns a Buffer).
const jpeg = await lx.sessions.frame("sess_abc123");
Quickstart — MCP server
// ~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"luxxon": {
"command": "npx",
"args": ["-y", "--package=@luxxon/mcp", "luxxon-mcp"],
"env": { "LUXXON_API_KEY": "lxxn_live_..." }
}
}
}
Restart Claude Desktop; the three tools appear under "Luxxon".
Status
v0, building in public. The SDK + MCP cover the agent-readsurface today (frames, session state, cancel). Session creationlands behind a custodial-wallet sign path that's still beingbuilt on the API side — for now you create sessions yourself viathe REST API and use this SDK/MCP to consume them.
Public dev console + Python/Go/CLI SDKs follow. Watch this repo.