BotIndex MCP Server — Market Intelligence Layer for AI Trading Agents #1

BotIndex MCP Server — Market Intelligence Layer for AI Trading Agents

Problem

AI trading agents using Bitget Agent Hub have execution capabilities (place orders, manage positions, query balances) but lack a structured market intelligence feed to inform what to trade and when.

Agents currently need to:

• Scrape multiple data sources for funding rate arbitrage signals
• Monitor token launch quality across platforms (Pump.fun, Doppler, Jupiter)
• Track cross-exchange price divergences
• Assess on-chain sentiment and whale flows

This is undifferentiated work that every agent builder reimplements.

Proposal: BotIndex as the Intelligence Layer

BotIndex MCP Server provides 11 structured endpoints covering crypto market intelligence — designed specifically for AI agent consumption.

Relevant endpoints for Bitget trading agents:

Integration Example

An agent with both MCP servers installed:

# codex.toml — or equivalent MCP config
[[mcp_servers]]
name = "bitget"
command = "npx"
args = ["-y", "bitget-mcp-server"]

[mcp_servers.env]
BITGET_API_KEY = "..."
BITGET_SECRET_KEY = "..."
BITGET_PASSPHRASE = "..."

[[mcp_servers]]
name = "botindex"
command = "npx"
args = ["-y", "botindex-mcp-server"]

Agent workflow:

1. botindex_funding_arb → finds BTC funding rate divergence (Bitget -0.03% vs HyperLiquid +0.05%)
2. bitget_futures_get_funding_rate → confirms Bitget's current rate
3. bitget_spot_place_order + bitget_futures_place_order → executes delta-neutral position
4. botindex_signals → monitors for exit signal

Details

• npm: botindex-mcp-server (1.3.1)
• Anthropic MCP Registry: io.github.Cyberweasel777/botindex-mcp-server
• GitHub: Cyberweasel777/botindex-mcp-server
• Free tier: 20 requests/day per endpoint, no API key required
• Spec: Agent Action Receipt (AAR) — every response includes a cryptographically signed receipt for agent-to-agent trust

Happy to collaborate on a combined quickstart guide or example agent that demonstrates the intelligence → execution pipeline.