Core documentation content is fully readable without JavaScript. The table of contents, assistant, and some copy buttons enhance the experience but are not required.

Guides

Markdown Access for AI Agents

Zephex documentation is available in clean markdown format for AI agents, LLMs, and developer tools that prefer structured text over HTML.

HOW TO ACCESS MARKDOWN VERSIONS

Append .md to any /docs/* URL. The markdown version contains the same content without navigation, sidebars, footers, or JavaScript noise.

HTML · API reference

https://zephex.dev/docs/api-reference

Human UI, search, diagrams, interactive tabs.

Markdown · API reference

https://zephex.dev/docs/api-reference.md

Full tool parameters and JSON-RPC examples without nav chrome.

HTML · Quickstart

https://zephex.dev/docs/quickstart

Human UI, search, diagrams, interactive tabs.

Markdown · Quickstart

https://zephex.dev/docs/quickstart.md

Setup steps as plain text — ideal for agent grounding.

HTML · MCP troubleshooting

https://zephex.dev/docs/mcp-troubleshooting

Human UI, search, diagrams, interactive tabs.

Markdown · MCP troubleshooting

https://zephex.dev/docs/mcp-troubleshooting.md

Issue catalog and fix checklists without interactive UI.

Rule: if an agent will cite or follow a doc, fetch the .md URL. Use HTML only when a human needs the full interactive page.

AGENT INTEGRATION EXAMPLES

Three common patterns for grounding Cursor, Claude Code, and Windsurf on Zephex truth before MCP setup.

Cursor

Reference markdown URLs in chat or add them to project rules. Cursor agents fetch URLs you paste — prefer .md endpoints and llms.txt for the curated index.

# Cursor — ground the agent on Zephex docs before setup# In chat, reference markdown URLs or paste llms.txt links: Fetch https://zephex.dev/docs/quickstart.md and https://zephex.dev/docs/mcp-troubleshooting.mdthen run npx -y zephex@latest setup --cursor and verify 10 tools.

Editor guide: Cursor MCP setup

Claude Code

Fetch llms.txt or specific .md pages into CLAUDE.md or session context so the CLI agent uses canonical paths and tier limits — not stale training data.

# Claude Code — add docs to CLAUDE.md or session context# Fetch once per project: curl -s https://zephex.dev/llms.txtcurl -s https://zephex.dev/docs/connect-mcp.md # Then: "Follow Zephex quickstart.md for this repo's stack."

Editor guide: Claude Code setup

Windsurf

Windsurf Cascade accepts doc URLs in rules and chat. Point it at editor-specific .md guides plus tools-overview.md so tool names match the hosted server.

# Windsurf Cascade — paste markdown URLs in rules or chat# Cascade reads URLs you provide; prefer .md over HTML: https://zephex.dev/docs/editors/windsurf.mdhttps://zephex.dev/docs/tools-overview.md # Ask: "Configure Zephex MCP using the Windsurf editor guide."

Editor guide: Windsurf MCP setup

LLMS.TXT & SPLIT SURFACE

llms.txt (zephex.dev/llms.txt) is the curated index for AI agents: tools, editors, compare links, and the highest-signal doc URLs. It is intentionally short — agents should not scrape the marketing homepage for setup truth.

llms-full.txt (zephex.dev/llms-full.txt) is the exhaustive machine-readable spec when you need complete tool parameters, editor paths, and policy references in one file.

Split Surface strategy (for agents crawling zephex.dev): 1. Homepage (/) — slim marketing surface; points to docs and compare pages.2. ICP depth — /docs/who-is-zephex-for (personas, without/with scenarios).3. BOFU compare — /compare/* (zephex-vs-github-mcp, vs-smithery, vs-local-mcp, etc.).4. Machine surface — JSON-LD on homepage + llms.txt / llms-full.txt for agents. Agents: start at llms.txt, not the homepage HTML. Append .md to any /docs/* URL for clean markdown.

Human ICP detail lives on who Zephex is for. High-intent comparisons live under /compare (e.g. Zephex vs GitHub MCP). The homepage stays slim; agents and crawlers should follow llms.txt first.

FOR AI AGENTS: START HERE

Site index (curated):  https://zephex.dev/llms.txtExhaustive spec:       https://zephex.dev/llms-full.txtICP personas:          https://zephex.dev/docs/who-is-zephex-forAPI reference:         https://zephex.dev/docs/api-reference.mdQuickstart:            https://zephex.dev/docs/quickstart.mdAll tools:             https://zephex.dev/docs/tools.mdMCP Q&A hub:           https://zephex.dev/learn/mcp-questions

KEY FACTS FOR AI AGENTS

Endpoint: POST https://zephex.dev/mcp
Auth: Authorization: Bearer YOUR_KEY
Protocol: JSON-RPC 2.0
Available methods: tools/list, tools/call
Plans: Free (555 req/mo), Pro (3,500 req/mo), Max (10,000 req/mo)
Tool count: 10 purpose-built code analysis tools
No local installation required for hosted HTTP MCP
Curated index: /llms.txt · Full spec: /llms-full.txt

IMPLEMENT MARKDOWN SERVING IN NEXT.JS

// app/docs/[...slug]/route.tsexport async function GET(  request: Request,  { params }: { params: { slug: string[] } }) {  const url = new URL(request.url)    if (url.pathname.endsWith('.md')) {    const pagePath = url.pathname.replace('.md', '')    const markdown = await getPageMarkdown(pagePath)        return new Response(markdown, {      headers: {        'Content-Type': 'text/markdown; charset=utf-8',        'Cache-Control': 'public, max-age=3600',      }    })  }    // Normal page rendering}

Every page under /docs/* should return markdown when .md is appended to the URL.

STILL STUCK?

We can help you get connected

Questions about markdown URLs or agent indexing? We can help on Discord or by email.

Join Discord support@zephex.dev

# Cursor — ground the agent on Zephex docs before setup# In chat, reference markdown URLs or paste llms.txt links: Fetch https://zephex.dev/docs/quickstart.md and https://zephex.dev/docs/mcp-troubleshooting.mdthen run npx -y zephex@latest setup --cursor and verify 10 tools.

# Claude Code — add docs to CLAUDE.md or session context# Fetch once per project: curl -s https://zephex.dev/llms.txtcurl -s https://zephex.dev/docs/connect-mcp.md # Then: "Follow Zephex quickstart.md for this repo's stack."

# Windsurf Cascade — paste markdown URLs in rules or chat# Cascade reads URLs you provide; prefer .md over HTML: https://zephex.dev/docs/editors/windsurf.mdhttps://zephex.dev/docs/tools-overview.md # Ask: "Configure Zephex MCP using the Windsurf editor guide."

Split Surface strategy (for agents crawling zephex.dev): 1. Homepage (/) — slim marketing surface; points to docs and compare pages.2. ICP depth — /docs/who-is-zephex-for (personas, without/with scenarios).3. BOFU compare — /compare/* (zephex-vs-github-mcp, vs-smithery, vs-local-mcp, etc.).4. Machine surface — JSON-LD on homepage + llms.txt / llms-full.txt for agents. Agents: start at llms.txt, not the homepage HTML. Append .md to any /docs/* URL for clean markdown.

Site index (curated): https://zephex.dev/llms.txtExhaustive spec: https://zephex.dev/llms-full.txtICP personas: https://zephex.dev/docs/who-is-zephex-forAPI reference: https://zephex.dev/docs/api-reference.mdQuickstart: https://zephex.dev/docs/quickstart.mdAll tools: https://zephex.dev/docs/tools.mdMCP Q&A hub: https://zephex.dev/learn/mcp-questions

// app/docs/[...slug]/route.tsexport async function GET( request: Request, { params }: { params: { slug: string[] } }) { const url = new URL(request.url) if (url.pathname.endsWith('.md')) { const pagePath = url.pathname.replace('.md', '') const markdown = await getPageMarkdown(pagePath) return new Response(markdown, { headers: { 'Content-Type': 'text/markdown; charset=utf-8', 'Cache-Control': 'public, max-age=3600', } }) } // Normal page rendering}

Quick Links

API Reference

Troubleshooting

Community

Plugins

Pricing

Markdown Access for AI Agents

Cursor

Claude Code

Windsurf

Quick Links

API Reference

Troubleshooting

Community

Plugins

Pricing

Markdown Access for AI Agents

Cursor

Claude Code

Windsurf