Skip to content
HanziDrill logo

HanziDrill MCP server

Which MCP servers are HanziDrill's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-11-16

← Back to HanziDrill

Whose server it is
  • Registry namecom.hanzidrill/hanzidrillverified in the MCP registry

The registry name sits under hanzidrill.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

When we asked itchecked 2026-11-16
  • Stateanswerschecked 2026-11-16
  • Tools8
  • Server calls itselfhanzidrill
  • Server version1.0.0
  • Protocol version2025-06-18

On 2026-11-16 the server answered initialize and tools/list without a sign-in. First answer in 1127 ms.

Should an agent connect it

Not proven either way yet
  • Whether its tools do the product's job has not been judged yet.
  • Checked on 5 of the 7 days it takes to judge reliability.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured daily

daily checks: 5, the first on 2026-11-16
  • INITE MCP scoreafter a weeka score needs 7 daily checks: 5 so far
  • Answeredmeasuring5 of the 7 daily checks it takes to print a share
  • Median first answer1019 ms
  • Space in an agent's context≈ 1,310 tokensthe definitions of all its tools (8), which an agent carries on every turn
  • Tool definitionsA · 98 of 100
  • Typed results0 of 8 toolsdeclared outputSchema: results come typed, read once
  • Parameters without a description: in 1 of 8

Worth a look

None of the five warning signs in these definitions.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers8 tools · read 2026-11-20
  • lookup_word

    Look up one Chinese word from the HSK 3.0 syllabus: pinyin, level, part of speech, English meaning, an example sentence, and how its level changed from HSK 2.0. Use when the user asks what a word means or which HSK level it is.

  • search_words

    Search HSK 3.0 words by level, topic tag, or pinyin (with or without tone marks or tone numbers). Returns at most 20 short matches per page. Give at least one of level, topic or pinyin.

  • get_character

    Get one Chinese character: pinyin, radical, stroke count, components (when known), its HSK character and writing levels, and common HSK words that use it. Includes a link to a stroke-order animation page when one exists.

  • hsk_level_wordlist

    List the words of one HSK 3.0 level in syllabus order, 50 per page, with pinyin and a short English meaning. Use for study lists; for the full list point the user to the download link.

  • compare_hsk2_hsk3

    Explain how a word or a whole level changed between the HSK 2.0 and HSK 3.0 syllabi: new, moved up, moved down, unchanged, or dropped, with counts for a level. Give either word or level.

  • practice_quiz

    Create a short HSK vocabulary quiz (1-10 multiple-choice questions) for a level: meaning, pinyin, fill-in-the-blank, or mixed. Answers are not included; show the questions to the user, collect their choices, then call grade_answers with…

  • grade_answers

    Grade a quiz created by practice_quiz. Returns which answers were right, the correct option, and a one-line explanation per question. This is practice feedback, not an exam score.

  • exam_dates

    List upcoming HSK test dates with registration deadlines and result dates, each with a link to its source notice. Dates can change; tell the user to confirm with their test center.

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "hanzidrill": {
      "type": "http",
      "url": "https://mcp.hanzidrill.com/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http hanzidrill https://mcp.hanzidrill.com/mcp

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://mcp.hanzidrill.com/mcp as a custom connector in the settings.

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.