Home · Learn · MCP integration

PredictAsiaX MCP for Claude — integration guide

Published 2026-09-05Reading time: 6 minCategory: Developer guide

TL;DR

The PredictAsiaX MCP server lets Claude Desktop and Claude Code directly search, analyze, and (with API key + user confirmation) trade event contracts on PredictAsiaX. Seven tools cover market search, probability tracking, portfolio read, order placement/cancellation, resolution evidence, and sandbox examples. Configuration takes under 5 minutes with a single JSON edit. Read-only tools are public; write tools require a Builder-Program API key with HMAC signing.

5-minute quick start

The fastest way to connect Claude Desktop to PredictAsiaX:

1. Edit your Claude Desktop config

// macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
// Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "pax": {
      "type": "sse",
      "url": "https://mcp.predictasiax.com/mcp/sse"
    }
  }
}

2. Restart Claude Desktop

The pax tools appear in the tools drawer within 5 seconds of restart.

3. Try it

Ask Claude: "Show me the top 5 prediction markets by probability movement in the last 24 hours." Claude will invoke pax.probability_movers and render the results.

Available tools

Tool	Auth	Description
`pax.market_search`	Public	Search markets by keyword, category, catalyst, or date range.
`pax.probability_movers`	Public	Top movers by probability change over configurable time window.
`pax.portfolio_read`	API key	Read user's open positions, filled orders, unrealized P&L.
`pax.place_order`	API key	Place an order (requires explicit user confirmation in Claude).
`pax.cancel_order`	API key	Cancel a specific order by order ID.
`pax.resolution_evidence`	Public	Fetch structured resolution evidence for settled markets.
`pax.sandbox_agent_examples`	Public	Runnable example flows for MCP-based prediction market agents.

Adding API key auth

For write tools (portfolio_read, place_order, cancel_order), you need an API key from the Builder Portal at builders.predictasiax.com. Once you have a key, add it to your MCP config:

{
  "mcpServers": {
    "pax": {
      "type": "sse",
      "url": "https://mcp.predictasiax.com/mcp/sse",
      "headers": {
        "X-Api-Key": "pax_live_xxxxxxxxxxxxxxxxxxxx"
      }
    }
  }
}

API keys are HMAC-signed at the server side. Full auth details at docs.predictasiax.com/auth.

Using from Claude Code (CLI)

Claude Code supports the same config format. Place at ~/.config/claude/mcp-config.json or pass with --mcp-config. Tools become available in interactive sessions and non-interactive (-p) mode.

Alternative: direct SDK use

If you're not using MCP, the official SDKs wrap the REST + WebSocket API:

# Python
pip install predictasiax

from predictasiax import Client
client = Client(api_key="pax_live_...")
markets = client.markets.search(keyword="election")

# TypeScript
npm install @predictasiax/sdk

import { PaxClient } from "@predictasiax/sdk";
const client = new PaxClient({ apiKey: "pax_live_..." });
const markets = await client.markets.search({ keyword: "election" });

Common issues

Tools not appearing after restart. Check JSON is valid (missing commas are the #1 cause). Restart Claude Desktop fully (quit + relaunch, not just close window).
429 Rate Limit. Public tools: 60 req/min per IP. Authenticated: 300 req/min per API key. Batch requests where possible.
401 Unauthorized on write tools. API key wrong or expired. Regenerate at builders.predictasiax.com.

Frequently asked questions

What is MCP (Model Context Protocol)?

The Model Context Protocol (MCP) is an open standard released by Anthropic in November 2024 that lets AI assistants (Claude Desktop, Claude Code, and any MCP-compatible client) connect to external data sources and tools. MCP servers expose tools, resources, and prompts over stdio or HTTP transport. Once configured, the AI can invoke the server's tools directly during conversation.

What can Claude do with the PredictAsiaX MCP server?

Once connected, Claude can: (1) search live prediction markets by keyword, category, or catalyst; (2) fetch top probability movers over configurable time windows; (3) read a user's portfolio positions (with API key); (4) place and cancel orders (with API key + explicit user confirmation); (5) retrieve resolution evidence for settled markets; (6) run sandbox agent example flows for development. All seven tools are documented at mcp.predictasiax.com/mcp/manifest.json.

How do I add PredictAsiaX to Claude Desktop?

Add the following to your Claude Desktop config at ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows). Include an 'mcpServers' object with a pax entry pointing to https://mcp.predictasiax.com/mcp/sse for SSE transport, or use the HTTP endpoint /mcp/call for stateless invocation. Restart Claude Desktop and the pax tools appear in the tools list.

Does the MCP server require authentication?

Read-only tools (market_search, probability_movers, resolution_evidence, sandbox_agent_examples) are public and require no authentication. Write tools (portfolio_read, place_order, cancel_order) require an API key issued from the PredictAsiaX Builders portal at builders.predictasiax.com. API keys are passed via the X-Api-Key header and are HMAC-signed on request. See docs.predictasiax.com/auth for full auth flow.

Can I use the MCP server from Claude Code?

Yes. Claude Code (the CLI) supports the same claude_desktop_config.json format via ~/.config/claude/mcp-config.json (or the CLI's --mcp-config flag). Once configured the MCP tools are available via the Skill and Bash tool wrappers in interactive and non-interactive Claude Code sessions.

Which SDK should I use for direct API integration?

PredictAsiaX provides official Python and TypeScript SDKs that wrap the REST + WebSocket API with HMAC signing built in. Python: pip install predictasiax; TypeScript: npm install @predictasiax/sdk. Both are open-source and Polymarket-compatible for developers migrating from Polymarket's API. Source and examples at github.com/predictasiax (currently under appeal for account restoration).

What's the difference between the MCP server and the REST API?

The REST API (docs.predictasiax.com/api) is a general HTTP interface for any client. The MCP server is a thin adapter on top of the REST API optimized for AI agents — it exposes the same operations as MCP tools with descriptions and JSON schemas that AI models can reason about, and it handles auth, rate limits, and idempotency headers automatically. For AI-driven workflows, use MCP; for traditional backend integration, use REST directly.

Is there a rate limit on the MCP server?

Yes. Public tools: 60 requests per minute per IP. Authenticated tools: 300 requests per minute per API key, higher tiers available for Builder Program participants. When rate-limited the response returns 429 with a Retry-After header. See docs.predictasiax.com/security for full rate limit policy.