Skip to main content

Documentation Index

Fetch the complete documentation index at: https://dev-docs.multihopper.com/llms.txt

Use this file to discover all available pages before exploring further.

MultiHopper’s documentation is available as a hosted Model Context Protocol (MCP) server. Connect any MCP-compatible AI client to search docs, read API references, and explore code examples in real time — without leaving your editor or agent workflow.

​
Server URL

https://dev-docs.multihopper.com/mcp
Paste this URL into any MCP-compatible client to connect.

​
Connect to your AI client

Open ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) and add:
{
  "mcpServers": {
    "multihopper-docs": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://dev-docs.multihopper.com/mcp"]
    }
  }
}
Restart Claude Desktop. The search_multi_hopper and query_docs_filesystem_multi_hopper tools will be available in every conversation.
Open Cursor Settings → MCP and add a new server:
  • Name: multihopper-docs
  • URL: https://dev-docs.multihopper.com/mcp
Or run the Mintlify install command in your terminal:
npx mcp add dev-docs.multihopper.com
Add to your .vscode/mcp.json or user settings:
{
  "servers": {
    "multihopper-docs": {
      "type": "sse",
      "url": "https://dev-docs.multihopper.com/mcp"
    }
  }
}
The server uses the standard SSE transport. Point your client at:
https://dev-docs.multihopper.com/mcp
No authentication is required for public documentation pages.

​
Available tools

​
search_multi_hopper

Searches across the MultiHopper knowledge base to find relevant information, code examples, API references, and guides. Use this tool for broad or conceptual queries — e.g. “how does the orchestrator work”, “what is keeperFundingTx”, or “error MH_039”. Returns contextual snippets with titles and links to the source pages. Example prompts that trigger this tool:
  • “How do I estimate fees before creating a transfer?”
  • “What is the broadcast order for prepared transactions?”
  • “What does phase recoverable mean?”

​
query_docs_filesystem_multi_hopper

Runs read-only shell-style queries against a virtualized filesystem containing all MultiHopper documentation pages and the OpenAPI spec. Use this tool when you need exact content from a specific page, regex matches across the docs, or structural exploration. Supported commands: rg, grep, find, tree, ls, cat, head, tail, stat, wc, sort, uniq, cut, sed, awk, jq Useful commands: 
# See the full docs structure
tree / -L 2

# Read a specific page
head -100 /quickstart.mdx
cat /api-reference/transfers/prepare.mdx

# Search for a keyword across all docs
rg -il "keeperFundingTx" /

# Show lines around a match with context
rg -C 3 "MH_039" /api-reference/

# List all OpenAPI endpoints
cat /api-reference/openapi.json | jq '.paths | keys'

# Read multiple pages in one call
head -80 /quickstart.mdx /guides/agentic-integration.mdx
Each call is stateless — the working directory always resets to /. Chain commands with && or use absolute paths if you need to navigate a subdirectory.

​
Example: agent workflow

Here’s how an AI agent would use both tools to answer “how do I sign and broadcast a prepared transfer?”:
  1. search_multi_hopper — "sign broadcast prepared transactions" → finds /guides/agentic-integration and /api-reference/transfers/prepare
  2. query_docs_filesystem_multi_hopper — head -150 /guides/agentic-integration.mdx → reads the full signing code examples
The search tool narrows scope; the filesystem tool reads the full page. Use them together.