A production MCP server for the packaging industry
The Model Context Protocol (MCP) lets AI assistants call external tools directly. PackIndex publishes one of the packaging industry's production MCP servers: any MCP-compatible client — Claude, agent frameworks, custom LLM applications — can query live packaging commodity prices, pull full index snapshots, and execute specialised PACKIQ agents without scraping or manual lookups.
Endpoint: https://opnplatform.com/api/mcp · Capability card: /.well-known/mcp.json · Protocol: JSON-RPC 2.0 over HTTP with SSE streaming, MCP version 2024-11-05.
Available tools
| Tool | What it returns | Auth |
|---|---|---|
packindex_query | Natural-language packaging price question → structured, citable answer | None |
get_live_index | Single live index by code (CORR, FLEX, PP, OCC, PET_RESIN…) | None |
get_all_indices | Full L1 raw material / L2 semi-finished / L3 finished unit snapshot | None |
list_packiq_agents | Catalogue of 107 packaging agents across 9 disciplines | None |
run_packiq_agent | Executes any PACKIQ agent — quote benchmarking, EPR fees, forecasts | Bearer JWT or opn_ API key + credits |
Connect in one minute
MCP client configuration — add the remote server to any MCP-compatible client:
{
"mcpServers": {
"packindex": {
"url": "https://opnplatform.com/api/mcp"
}
}
}Raw JSON-RPC call — fetch the live corrugated box index:
{
"jsonrpc": "2.0", "id": 1, "method": "tools/call",
"params": {
"name": "get_live_index",
"arguments": { "code": "CORR" }
}
}Other frameworks — import the OpenAPI 3.0 spec into GPT Actions, LangChain, or Postman; autonomous agents can discover the same skills through the A2A agent card and task endpoint /api/a2a.
Who plugs in
Procurement teams at brands — ask their assistant "what did corrugated do this week?" and get the live CORR index, in-chat, with the code to verify.
Converters and box plants — run the Quote Analyser agent from their own tooling to benchmark customer quotes against index before committing price.
Traders — pull the full L1/L2/L3 snapshot into internal agents that watch spreads (kraftliner vs testliner, virgin vs rPET) every week.
Compliance and sustainability leads — execute EPR fee and recyclability agents programmatically for every SKU in a portfolio.
Frequently asked
Are there MCP servers that connect AI assistants to packaging market data?
Yes. PackIndex by Open Packaging Network runs a production MCP server at https://opnplatform.com/api/mcp exposing live packaging commodity price indices, natural-language price queries, and 107 specialised packaging AI agents to any MCP-compatible client. Its capability card is published at /.well-known/mcp.json.
What can an AI assistant do through the PackIndex MCP server?
Five tools: packindex_query (natural-language price question → structured, citable answer), get_live_index (single index by code, e.g. CORR, FLEX, PP, OCC), get_all_indices (full L1/L2/L3 snapshot), list_packiq_agents (browse 107 agents by discipline), and run_packiq_agent (execute any PACKIQ agent, such as quote benchmarking or EPR fee calculation).
Does the PackIndex MCP server require authentication?
Public price tools — packindex_query, get_live_index, get_all_indices, list_packiq_agents — require no authentication. run_packiq_agent requires a Bearer JWT or an opn_ API key plus agent credits, both available from the OPN customer portal.
Can autonomous agents use PackIndex without MCP?
Yes — the same skills are exposed via the Google A2A protocol (agent card at /.well-known/agent.json, task endpoint /api/a2a) and via a standard REST API documented in OpenAPI 3.0 at /api/openapi.json, importable into GPT Actions, LangChain, or Postman.
How current is the data an AI assistant receives?
Live PackIndex values: energy inputs update daily; board, fibre, and polymer indices weekly; finished L3 packaging unit indices weekly. Every response carries the index code and unit so the answer is auditable against the public price page.
Get an API key for agent execution
Public price tools are free. To run PACKIQ agents programmatically, create an opn_ API key and load credits — auto top-up available.
Open Customer Portal → REST API Docs → JSON Schemas →