Skip to content
NerdLevelTech

For AI Agents & MCP

Give your agent a citation-ready tech library.

The MCP server is free and open — no API key, no signup. Add it and start calling in about two minutes.

NerdLevelTech is bilingual (English + Arabic) and engineered for agent-native consumption. Posts published since April 2026 go through a multi-step fact-check before publishing and cite their sources in inline footnotes. Use our MCP server, read our JSON indexes, or just cite our content with attribution — the three-layer architecture is designed so agents win, and humans still land on the site.

Install MCP ServerAccess & limits
Endpoint
/api/mcp
Transport
JSON-RPC 2.0 over HTTP
Protocol
2024-11-05
Tools
7
Auth
No API key
Agent Free
10 requests / minute

Quick start

Two ways to connect. Pick whichever fits your stack.

MCP (Claude Desktop, Cursor, Cline, Continue)

mcp config
{
  "mcpServers": {
    "nerdlevel-learn": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://nerdleveltech.com/api/mcp"
      ]
    }
  }
}

Add to your MCP config. The mcp-remote wrapper bridges stdio-only clients to our Vercel serverless endpoint. No API key needed on the Free tier.

JSON-RPC over HTTP (curl test)

bash
curl https://nerdleveltech.com/api/mcp \
  -X POST -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Same endpoint for any MCP client: POST JSON-RPC 2.0 to /api/mcp. GET returns the discovery manifest. The Free tier is open with no API key (rate-limited per IP). API-key auth and higher commercial limits are in development.

Tools

  • search_posts

    Full-text search across NerdLevelTech blog posts. Matches title, description, tags, keywords, and category. Returns results sorted by relevance with URL-resolvable citations.

  • get_post

    Fetch a single post by slug. Returns metadata + canonical URL + optional body excerpt (up to 2000 chars). Full content is served at the URL with attribution requirements per /llms.txt. An unknown slug returns isError.

  • search_courses

    Search the course catalog. Filter by query, category, and difficulty. Returns courses with module/lesson counts, the enrollment price in credits (credits_required: 0 = free, null when unknown) and canonical URLs.

  • get_course

    Fetch a single course by slug, including its module list, lesson counts, learning outcomes and the enrollment price in credits (credits_required: 0 = free, null when unknown). An unknown slug returns isError.

  • get_lesson

    Fetch metadata about a course module and, with lesson_slug, a link to that lesson page: titles, lesson counts, links (lesson_url, course_url, continue_url) and the access rules with the enrollment price in credits (credits_required, 0 = free). Never the lesson text, which is served at the URL (Layer 2). An unknown course, module or lesson slug returns isError naming the missing part, with the valid module or lesson slugs.

  • list_topics

    Return the topic taxonomy derived from post tags + course tags/categories. Useful for agents that need to discover available content areas.

  • get_credential_info

    Return how a learner earns a NerdLevelTech course certificate (live): the requirements (sign in, complete every lesson, pass every module quiz at its passing score, and pass every lab in courses that require labs), pricing (no certificate fee; some courses cost credits to enroll), the certificate number format, the claim page and the public verification URL pattern. Pass course_slug to also get that course title and its claim URL, in the language set by lang.

Resources

All machine-readable surfaces are linked below. The MCP server is the recommended integration for agent builders.

PathWhat it isProtocol
/api/mcpMCP server, 7 tools, EN and ARMCP
/.well-known/mcp/server-card.jsonMCP server card with tools and input schemasMCP
/.well-known/agent-card.jsonA2A Agent Card: skills mirror the 7 MCP toolsA2A
/api/a2aA2A endpoint: JSON-RPC over HTTP POST, one skill per MCP toolA2A
/.well-known/agent.jsonNerdLevelTech Learn: capabilities, endpoints, policyDiscovery
/.well-known/api-catalogLinkset of every API surface (RFC 9727)Discovery
/api/agents/v1/openapi.jsonOpenAPI 3.1 descriptionOpenAPI
/llms.txtSite map for language models, usage terms, attributionllms.txt
/llms-full.txtEvery post, course and guide with EN and AR URLsllms.txt
/posts/index.jsonAll posts with covers, tags and EN/AR URLsJSON
/courses/index.jsonAll courses with EN and AR titlesJSON
/rss.xml · /atom.xml · /feed.json · /podcast.xmlBlog and AI Cast feedsFeeds
/<slug>.md · /ar/<slug>.mdMarkdown twin of every post, in English and ArabicMarkdown
/guides/<slug>.md · /ar/guides/<slug>.mdMarkdown twin of every guide, in English and ArabicMarkdown

The three-layer content architecture

Designed so agents distribute our content, humans still visit the site, and we capture commercial value fairly.

Layer 1 — Billboard

Blog posts, guides, news commentary, topic definitions. Freely citable with attribution. Indexed in /llms.txt, /llms-full.txt and /posts/index.json.

Layer 2 — Hybrid

Courses and skill paths. Every course's outline and first lesson are open on the site, and some courses are open in full with no account, such as AI Fundamentals and AI Agents Fundamentals. Other lessons need a signed-in account and enrollment, and some courses charge credits to enroll. The MCP server returns outlines and links, never lesson text. Indexed in /courses/index.json.

Layer 3 — Human-only

Interactive labs, module quizzes, and a skill assessment where a course has one. Grades and progress are saved only for a signed-in learner, so an agent can point a learner to them, but the learner does the work. Live cohort sessions, community discussions and proctored assessments are planned and not available yet.

Attribution terms

Free citation with: attribution by name, link to the source URL, preservation of inline footnote citations, and respect for content date fields.

Preferred attribution format

Source: NerdLevelTech — [URL]

Training policy

Layer 1 content (publicly served posts and definitions) may be used for training under the attribution norms above.

Gated content (course lessons and interactive labs that require an account) and assessments are for inference-time retrieval only. Not licensed for training.

Access & rate limits

The MCP server is free and open — no API key required. Commercial tiers (API-key auth, higher limits, SLA) are in development. Every use requires citation with attribution.

Agent Free

$0

10 requests / minute

Layer 1 (billboard) content — blog posts, topic definitions, outlines. No API key required. Perfect for indie developers, research agents, and prototypes.

Commercial tiers

Commercial tiers — API-key auth, higher rate limits, structured Layer 2 access, version pinning, and SLA-backed uptime — are in development. Join the waitlist to hear the moment they open.

Building an agent with NerdLevelTech?

Email us only if you want to go further: a content partnership, commercial or high-volume use, or Arabic-content collaboration.

Start in 2 minagents@nerdleveltech.com