duden-mcp 0.1.0__tar.gz

duden_mcp-0.1.0/.github/workflows/publish.yml

@@ -0,0 +1,20 @@
+name: Publish to PyPI
+
+on:
+  push:
+    tags: ["v*"]
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    permissions:
+      id-token: write # required for PyPI trusted publishing (OIDC)
+      contents: read
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - run: pip install build
+      - run: python -m build
+      - uses: pypa/gh-action-pypi-publish@release/v1

duden_mcp-0.1.0/.github/workflows/registry.yml

@@ -0,0 +1,61 @@
+name: Publish to MCP Registry
+
+# Official MCP Registry (registry.modelcontextprotocol.io) via GitHub OIDC.
+# No secret, no interactive login: the runner's OIDC token proves ownership
+# of the io.github.mohammad-abbasi-sh/* namespace.
+# Refs: https://github.com/modelcontextprotocol/registry
+
+on:
+  push:
+    tags: ["v*"]
+  workflow_dispatch:
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    permissions:
+      id-token: write # required for OIDC -> MCP Registry
+      contents: read
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Sync server.json version to tag
+        # Only sync when running on a version tag (v*). On workflow_dispatch
+        # (ref = refs/heads/main) keep the version committed in server.json.
+        run: |
+          VER=""
+          if [[ "$GITHUB_REF" == refs/tags/* ]]; then
+            TAG="${GITHUB_REF#refs/tags/}"
+            VER="${TAG#v}"
+          fi
+          if [ -n "$VER" ]; then
+            python3 - "$VER" << 'EOF'
+          import json, sys
+          ver = sys.argv[1]
+          d = json.load(open('server.json'))
+          d['version'] = ver
+          for pkg in d.get('packages', []):
+              pkg['version'] = ver
+          json.dump(d, open('server.json', 'w'), indent=2, ensure_ascii=False)
+          print('server.json version ->', ver)
+          EOF
+          else
+            echo "No version tag ($GITHUB_REF); keeping server.json version as-is"
+          fi
+
+      - name: Install mcp-publisher
+        run: |
+          curl -sL "https://github.com/modelcontextprotocol/registry/releases/download/v1.8.7/mcp-publisher_linux_amd64.tar.gz" -o /tmp/mcp-publisher.tar.gz
+          tar -xzf /tmp/mcp-publisher.tar.gz -C /tmp
+          sudo mv /tmp/mcp-publisher /usr/local/bin/mcp-publisher
+          mcp-publisher --version
+
+      - name: Validate server.json
+        run: mcp-publisher validate server.json
+
+      - name: Publish to MCP Registry (OIDC)
+        run: |
+          mcp-publisher login github-oidc
+          mcp-publisher publish server.json
+        # A registry outage must never break the release; re-run from Actions if needed.
+        continue-on-error: true

duden_mcp-0.1.0/.github/workflows/smoke.yml

@@ -0,0 +1,21 @@
+name: Daily live smoke
+
+# One fixed probe per day: catches fingerprint rot / bot-wall changes
+# before users hit them. Fails visibly on 403/timeout.
+on:
+  schedule:
+    - cron: "0 6 * * *"
+  workflow_dispatch:
+
+jobs:
+  smoke:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Install
+        run: pip install -e .
+      - name: Smoke probe
+        run: python tests/smoke_live.py

duden_mcp-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Mohammad Abbasi
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

duden_mcp-0.1.0/PKG-INFO

@@ -0,0 +1,75 @@
+Metadata-Version: 2.5
+Name: duden-mcp
+Version: 0.1.0
+Summary: Unofficial read-only MCP server for Duden (duden.de) — dictionary lookup, synonyms, spelling & verb conjugation
+Author: Mohammad Abbasi
+License: MIT
+License-File: LICENSE
+Requires-Python: >=3.10
+Requires-Dist: httpx>=0.27
+Requires-Dist: mcp<2
+Description-Content-Type: text/markdown
+
+# duden-mcp 📖
+
+mcp-name: io.github.mohammad-abbasi-sh/duden-mcp
+
+Unofficial **read-only** MCP server for **[duden.de](https://www.duden.de)** — Duden dictionary lookup from duden.de (free, no key, no login): word meanings, synonyms, spelling with hyphenation, grammar notes and full verb conjugation. No key, no login.
+
+> Built by Mohammad Abbasi as part of the "Germany MCP factory". Not affiliated with duden.de.
+
+## For AI agents
+
+Add to your MCP client config (stdio):
+
+```json
+{
+  "mcpServers": {
+    "duden": {
+      "command": "uvx",
+      "args": ["duden-mcp==0.1.0"]
+    }
+  }
+}
+```
+
+Or run locally: `python -m duden_mcp.server` (needs `mcp<2`, `httpx`).
+
+## Tools
+
+| Tool | Description |
+|------|-------------|
+| `duden_lookup_word` | Look up a German word in Duden: meanings, word type, frequency, hyphenation & grammar — `word` (e.g. Haus, laufen) |
+| `duden_synonyms` | Synonyms of a German word from Duden — `word` |
+| `duden_spelling` | Spelling of a German word: hyphenation and spelling notes — `word` |
+| `duden_grammar` | Grammar notes for a German word: declension/conjugation summary — `word` |
+| `duden_conjugate` | Full conjugation table of a German verb: present, past, perfect, participles, imperative — `verb` |
+| `duden_search` | Search Duden headwords matching a query — `query`, `limit` 1..20 |
+| `duden_health` | Reachability probe (ok / blocked / error) — one polite request |
+
+## Notes
+
+- All data comes from **Duden's public word pages** (free, no key, no login).
+- Meanings are the numbered definitions from the word entry; synonyms are the ones listed on the entry (the dedicated `/synonyme/<word>` overview page only lists alphabetical navigation, so it is linked, not parsed).
+- Word URLs use Duden's transliterated slug (ä→ae, ö→oe, ü→ue, ß→ss); lookup tries the transliterated form first, then the raw spelling.
+- Ambiguous headwords redirect: `sein` → `sein_Hilfsverb`, `haben` → `haben_Hilfsverb` / `haben_Vollverb`. The tools follow Duden's own redirects, and `duden_conjugate` resolves the suffixed conjugation page through the word entry when the bare `/konjugation/<verb>` URL 404s.
+- Hyphenation marks syllable breaks with `·` (Duden itself writes a soft hyphen or pipe).
+- Conjugation tables include the reflexive pronouns Duden lists (e.g. `laufe (mich/mir)` for `laufen`).
+
+Tests: 37/37 green (28 offline + 9 live, 2026-11-10).
+
+## Development
+
+```bash
+python3 -m venv .venv && .venv/bin/pip install "mcp<2" httpx pytest
+.venv/bin/pip install -e .
+.venv/bin/python -m pytest tests/ -q
+```
+
+## License
+
+MIT
+
+---
+
+Built by [Mohammad Abbasi](https://github.com/mohammad-abbasi-sh) — [Almanibama](https://almanibama.ir) · [mabbasi.ir](https://mabbasi.ir)

duden_mcp-0.1.0/README-de.md

@@ -0,0 +1,52 @@
+# duden-mcp 📖
+
+mcp-name: io.github.mohammad-abbasi-sh/duden-mcp
+
+Inoffizieller **nur lesender** MCP-Server für **[duden.de](https://www.duden.de)** — Duden-Nachschlagen auf duden.de (kostenlos, kein Schlüssel, kein Login): Bedeutungen, Synonyme, Rechtschreibung mit Worttrennung, Grammatik und vollständige Verbkonjugation. Kein Schlüssel, keine Anmeldung.
+
+> Gebaut von Mohammad Abbasi als Teil der "Germany MCP factory". Nicht mit duden.de verbunden.
+
+## Für KI-Agenten
+
+Zur MCP-Client-Konfiguration hinzufügen (stdio):
+
+```json
+{
+  "mcpServers": {
+    "duden": {
+      "command": "uvx",
+      "args": ["duden-mcp==0.1.0"]
+    }
+  }
+}
+```
+
+Oder lokal starten: `python -m duden_mcp.server` (benötigt `mcp<2`, `httpx`).
+
+## Tools
+
+| Tool | Beschreibung |
+|------|-------------|
+| `duden_lookup_word` | Ein deutsches Wort im Duden nachschlagen: Bedeutungen, Wortart & Grammatik |
+| `duden_synonyms` | Synonyme eines deutschen Wortes |
+| `duden_spelling` | Rechtschreibung & Worttrennung eines deutschen Wortes |
+| `duden_grammar` | Grammatiknotizen zu einem deutschen Wort |
+| `duden_conjugate` | Vollständige Konjugationstabelle eines deutschen Verbs |
+| `duden_search` | Duden-Stichwörter suchen |
+| `duden_health` | Erreichbarkeitsprüfung (ok / blocked / error) — eine höfliche Anfrage |
+
+## Entwicklung
+
+```bash
+python3 -m venv .venv && .venv/bin/pip install "mcp<2" httpx pytest
+.venv/bin/pip install -e .
+.venv/bin/python -m pytest tests/ -q
+```
+
+## Lizenz
+
+MIT
+
+---
+
+Built by [Mohammad Abbasi](https://github.com/mohammad-abbasi-sh) — [Almanibama](https://almanibama.ir) · [mabbasi.ir](https://mabbasi.ir)

duden_mcp-0.1.0/README-fa.md

@@ -0,0 +1,52 @@
+# duden-mcp 📖
+
+mcp-name: io.github.mohammad-abbasi-sh/duden-mcp
+
+سرور MCP غیررسمی و **فقط‌خواندنی** برای **[duden.de](https://www.duden.de)** — جست‌وجو در فرهنگ لغت دودن از duden.de (رایگان، بدون کلید): معانی واژه، مترادف‌ها، املا با جداسازی هجاها، نکات دستوری و صرف کامل فعل. بدون کلید، بدون ورود.
+
+> ساختهٔ محمد عباسی، بخشی از «کارخانهٔ MCP آلمان». وابسته به duden.de نیست.
+
+## برای ایجنت‌های هوش مصنوعی
+
+به پیکربندی MCP کلاینت اضافه کنید (stdio):
+
+```json
+{
+  "mcpServers": {
+    "duden": {
+      "command": "uvx",
+      "args": ["duden-mcp==0.1.0"]
+    }
+  }
+}
+```
+
+یا به‌صورت محلی اجرا کنید: `python -m duden_mcp.server` (نیازمند `mcp<2` و `httpx`).
+
+## ابزارها
+
+| ابزار | توضیح |
+|------|-------|
+| `duden_lookup_word` | جست‌وجوی واژهٔ آلمانی در دودن: معانی، نوع کلمه و دستور زبان |
+| `duden_synonyms` | مترادف‌های یک واژهٔ آلمانی |
+| `duden_spelling` | املا و جداسازی هجاهای یک واژهٔ آلمانی |
+| `duden_grammar` | نکات دستوری یک واژهٔ آلمانی |
+| `duden_conjugate` | جدول کامل صرف یک فعل آلمانی |
+| `duden_search` | جست‌وجوی سرخط‌های دودن |
+| `duden_health` | بررسی دسترس‌پذیری (ok / blocked / error) — یک درخواست مؤدبانه |
+
+## توسعه
+
+```bash
+python3 -m venv .venv && .venv/bin/pip install "mcp<2" httpx pytest
+.venv/bin/pip install -e .
+.venv/bin/python -m pytest tests/ -q
+```
+
+## لایسنس
+
+MIT
+
+---
+
+Built by [Mohammad Abbasi](https://github.com/mohammad-abbasi-sh) — [Almanibama](https://almanibama.ir) · [mabbasi.ir](https://mabbasi.ir)

duden_mcp-0.1.0/README.md

@@ -0,0 +1,63 @@
+# duden-mcp 📖
+
+mcp-name: io.github.mohammad-abbasi-sh/duden-mcp
+
+Unofficial **read-only** MCP server for **[duden.de](https://www.duden.de)** — Duden dictionary lookup from duden.de (free, no key, no login): word meanings, synonyms, spelling with hyphenation, grammar notes and full verb conjugation. No key, no login.
+
+> Built by Mohammad Abbasi as part of the "Germany MCP factory". Not affiliated with duden.de.
+
+## For AI agents
+
+Add to your MCP client config (stdio):
+
+```json
+{
+  "mcpServers": {
+    "duden": {
+      "command": "uvx",
+      "args": ["duden-mcp==0.1.0"]
+    }
+  }
+}
+```
+
+Or run locally: `python -m duden_mcp.server` (needs `mcp<2`, `httpx`).
+
+## Tools
+
+| Tool | Description |
+|------|-------------|
+| `duden_lookup_word` | Look up a German word in Duden: meanings, word type, frequency, hyphenation & grammar — `word` (e.g. Haus, laufen) |
+| `duden_synonyms` | Synonyms of a German word from Duden — `word` |
+| `duden_spelling` | Spelling of a German word: hyphenation and spelling notes — `word` |
+| `duden_grammar` | Grammar notes for a German word: declension/conjugation summary — `word` |
+| `duden_conjugate` | Full conjugation table of a German verb: present, past, perfect, participles, imperative — `verb` |
+| `duden_search` | Search Duden headwords matching a query — `query`, `limit` 1..20 |
+| `duden_health` | Reachability probe (ok / blocked / error) — one polite request |
+
+## Notes
+
+- All data comes from **Duden's public word pages** (free, no key, no login).
+- Meanings are the numbered definitions from the word entry; synonyms are the ones listed on the entry (the dedicated `/synonyme/<word>` overview page only lists alphabetical navigation, so it is linked, not parsed).
+- Word URLs use Duden's transliterated slug (ä→ae, ö→oe, ü→ue, ß→ss); lookup tries the transliterated form first, then the raw spelling.
+- Ambiguous headwords redirect: `sein` → `sein_Hilfsverb`, `haben` → `haben_Hilfsverb` / `haben_Vollverb`. The tools follow Duden's own redirects, and `duden_conjugate` resolves the suffixed conjugation page through the word entry when the bare `/konjugation/<verb>` URL 404s.
+- Hyphenation marks syllable breaks with `·` (Duden itself writes a soft hyphen or pipe).
+- Conjugation tables include the reflexive pronouns Duden lists (e.g. `laufe (mich/mir)` for `laufen`).
+
+Tests: 37/37 green (28 offline + 9 live, 2026-11-10).
+
+## Development
+
+```bash
+python3 -m venv .venv && .venv/bin/pip install "mcp<2" httpx pytest
+.venv/bin/pip install -e .
+.venv/bin/python -m pytest tests/ -q
+```
+
+## License
+
+MIT
+
+---
+
+Built by [Mohammad Abbasi](https://github.com/mohammad-abbasi-sh) — [Almanibama](https://almanibama.ir) · [mabbasi.ir](https://mabbasi.ir)

duden_mcp-0.1.0/pyproject.toml

@@ -0,0 +1,25 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "duden-mcp"
+version = "0.1.0"
+description = "Unofficial read-only MCP server for Duden (duden.de) — dictionary lookup, synonyms, spelling & verb conjugation"
+readme = "README.md"
+license = { text = "MIT" }
+requires-python = ">=3.10"
+dependencies = [
+    "mcp<2",
+    "httpx>=0.27",
+]
+authors = [{ name = "Mohammad Abbasi" }]
+
+[project.scripts]
+duden-mcp = "duden_mcp.server:main"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/duden_mcp"]
+
+[tool.pytest.ini_options]
+markers = ["live: tests that hit the real https://www.duden.de site"]

duden_mcp-0.1.0/server.json

@@ -0,0 +1 @@
+{"$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json", "description": "Unofficial read-only MCP server for Duden — meanings, synonyms, spelling & conjugation.", "name": "io.github.mohammad-abbasi-sh/duden-mcp", "packages": [{"identifier": "duden-mcp", "registryBaseUrl": "https://pypi.org", "registryType": "pypi", "transport": {"type": "stdio"}, "version": "0.1.0"}], "repository": {"source": "github", "url": "https://github.com/mohammad-abbasi-sh/duden-mcp"}, "title": "Duden — Germany's standard dictionary: word meanings, synonyms, spelling & grammar", "version": "0.1.0", "websiteUrl": "https://github.com/mohammad-abbasi-sh/duden-mcp"}

duden_mcp-0.1.0/src/duden_mcp/__init__.py

@@ -0,0 +1,7 @@
+"""duden_mcp — unofficial read-only MCP server for duden.de.
+
+Duden dictionary lookup from duden.de (free, no key, no login): word meanings, synonyms, spelling with hyphenation, grammar notes and full verb conjugation.
+"""
+from __future__ import annotations
+
+__version__ = "0.1.0"

duden_mcp-0.1.0/src/duden_mcp/domain.py

@@ -0,0 +1,232 @@
+"""Domain logic: shape Duden word / conjugation / search pages into
+compact dicts (stdlib only).
+
+Data source: duden.de public word pages
+(https://www.duden.de/rechtschreibung/<word>) — free, no key, no login.
+
+Page structure (server-rendered HTML, verified 2026-11-10):
+- lemma in <h1>; word type / frequency in <dl class="tuple"> blocks
+  (Wortart, Häufigkeit with a shaft aria-label).
+- meanings: <div class="enumeration__text">…</div> inside the
+  #bedeutungen division (one per numbered meaning).
+- spelling: #rechtschreibung division (Worttrennung line + notes).
+- synonyms: links with data-duden-ref-type="lexeme" inside the
+  #synonyme division; the dedicated /synonyme/<word> overview page
+  only lists alphabetical navigation, so the word page's list is used.
+- grammar: first <p> of the #grammatik division
+  ("das Haus; Genitiv: des Hauses, Häuser"); verbs link to
+  /konjugation/<verb>.
+- conjugation page: <ul class="accordion__list content-column"> lists
+  in document order — first 6 are Indikativ (Präsens, Präteritum,
+  Perfekt, Plusquamperfekt, Futur I, Futur II), then Konjunktiv I/II,
+  then the Imperativ list (2 items, no label) and the infinite forms
+  (Infinitiv mit zu / Partizip I / Partizip II, one form each).
+  Footnote <sup> markers are stripped before parsing.
+"""
+from __future__ import annotations
+
+import html as _html
+import re
+from urllib.parse import quote
+
+BASE = "https://www.duden.de"
+
+PERSONS = ["ich", "du", "er/sie/es", "wir", "ihr", "sie"]
+
+TENSES = ["Präsens", "Präteritum", "Perfekt", "Plusquamperfekt",
+          "Futur I", "Futur II"]
+TENSE_KEYS = ["praesens", "praeteritum", "perfekt", "plusquamperfekt",
+              "futur_i", "futur_ii"]
+
+_TAG_RE = re.compile(r"<[^>]+>")
+_SUP_RE = re.compile(r"<sup[^>]*>.*?</sup>", re.S)
+_WS_RE = re.compile(r"\s+")
+
+_TRANSLIT = str.maketrans({
+    "ä": "ae", "ö": "oe", "ü": "ue",
+    "Ä": "Ae", "Ö": "Oe", "Ü": "Ue", "ß": "ss",
+})
+
+
+def validate_word(word: str) -> str:
+    w = (word or "").strip()
+    if not w:
+        raise ValueError("word must not be empty")
+    if len(w) > 80:
+        raise ValueError("word too long (max 80 characters)")
+    return w
+
+
+def validate_limit(limit: int) -> int:
+    if not 1 <= limit <= 20:
+        raise ValueError(f"limit out of range: {limit} (expected 1..20)")
+    return limit
+
+
+def word_slug(word: str) -> str:
+    """Duden URL slug: umlauts transliterated (ä→ae …), spaces → _."""
+    return word.translate(_TRANSLIT).replace(" ", "_")
+
+
+def word_url(word: str) -> str:
+    return f"{BASE}/rechtschreibung/{word_slug(word)}"
+
+
+def conjugation_url(verb: str) -> str:
+    return f"{BASE}/konjugation/{word_slug(verb)}"
+
+
+def search_url(query: str) -> str:
+    return f"{BASE}/suchen/dudenonline/{quote(query.strip())}"
+
+
+def _clean(text: str) -> str:
+    return _WS_RE.sub(" ", _html.unescape(_TAG_RE.sub(" ", text or ""))).strip()
+
+
+def _division(html: str, div_id: str) -> str:
+    """Raw HTML of the <div class="division" id="<div_id>"> block."""
+    m = re.search(rf'id="{re.escape(div_id)}"', html)
+    if not m:
+        return ""
+    start = html.rfind("<div", 0, m.start())
+    if start < 0:
+        start = m.start()
+    nxt = html.find('<div class="division ', start + 10)
+    return html[start:] if nxt < 0 else html[start:nxt]
+
+
+def _first_paragraph(division_html: str) -> str:
+    m = re.search(r"<p[^>]*>(.*?)</p>", division_html, re.S)
+    return _clean(m.group(1)) if m else ""
+
+
+def parse_word_page(html: str) -> dict:
+    """{lemma, word_type, frequency, meanings, hyphenation, grammar,
+    origin, synonyms, conjugation_url, url} — empty strings/lists
+    when a section is absent (not every word has every section)."""
+    m = re.search(r"<h1[^>]*>(.*?)</h1>", html, re.S)
+    lemma = _clean(m.group(1)) if m else ""
+    lemma = lemma.replace("­", "").replace(" ,", ",")
+    m = re.search(
+        r'<dt class="tuple__key">Wortart:.*?</dt>\s*'
+        r'<dd class="tuple__val">(.*?)</dd>', html, re.S)
+    word_type = _clean(m.group(1)) if m else ""
+    m = re.search(
+        r'<dt class="tuple__key">Häufigkeit:.*?'
+        r'<div class="shaft" aria-label="([^"]+)"', html, re.S)
+    frequency = _html.unescape(m.group(1)) if m else ""
+
+    meanings = []
+    for raw in re.findall(r'enumeration__text">(.*?)</div>', html, re.S):
+        t = _clean(raw)
+        if t and t not in meanings:
+            meanings.append(t)
+
+    rechtschreibung = _division(html, "rechtschreibung")
+    hyphenation = ""
+    if rechtschreibung:
+        lines = [_clean(ln) for ln in _TAG_RE.sub("\n", rechtschreibung).split("\n")]
+        lines = [ln for ln in lines if ln]
+        for i, ln in enumerate(lines):
+            if ln == "Worttrennung" and i + 1 < len(lines):
+                # Duden marks syllable breaks with a soft hyphen or
+                # a pipe; normalise both to · for display
+                hyphenation = (lines[i + 1].replace("­", "·")
+                               .replace("|", "·"))
+                break
+
+    grammar = _first_paragraph(_division(html, "grammatik"))
+    origin = _first_paragraph(_division(html, "herkunft"))
+    syn_div = _division(html, "synonyme")
+    synonyms = []
+    for raw in re.findall(
+            r'data-duden-ref-type="lexeme"[^>]*>([^<]+)</a>', syn_div):
+        w = _clean(raw)
+        if w and w not in synonyms:
+            synonyms.append(w)
+    m = re.search(r'href="(/konjugation/[^"]+)"', html)
+    conj_url = f"{BASE}{m.group(1)}" if m else ""
+    return {
+        "lemma": lemma,
+        "word_type": word_type,
+        "frequency": frequency,
+        "meanings": meanings,
+        "hyphenation": hyphenation,
+        "grammar": grammar,
+        "origin": origin,
+        "synonyms": synonyms,
+        "conjugation_url": conj_url,
+    }
+
+
+def parse_spelling(html: str) -> dict:
+    """{hyphenation, notes} from the #rechtschreibung division."""
+    word = parse_word_page(html)
+    div = _division(html, "rechtschreibung")
+    notes = _clean(div)[:1500]
+    # drop the leading "Rechtschreibung" heading + info icon
+    notes = re.sub(r"^Rechtschreibung\s*ⓘ\s*", "", notes)
+    return {"hyphenation": word["hyphenation"], "notes": notes}
+
+
+def parse_conjugation(html: str) -> dict:
+    """{indikativ: {tense_key: {person: form}}, imperative:
+    {singular, plural}, infinitiv_mit_zu, partizip_i, partizip_ii}."""
+    clean_html = _SUP_RE.sub("", html)
+    uls = re.findall(
+        r'<ul class="accordion__list content-column">(.*?)</ul>',
+        clean_html, re.S)
+    tables = []
+    for u in uls:
+        items = [_clean(li) for li in re.findall(
+            r'<li class="accordion__item">(.*?)</li>', u, re.S)]
+        items = [it for it in items if it]
+        if items:
+            tables.append(items)
+
+    indikativ: dict[str, dict[str, str]] = {}
+    imperative = {"singular": "", "plural": ""}
+    infinitiv_mit_zu = ""
+    partizip_i = ""
+    partizip_ii = ""
+    for items in tables:
+        label, forms = items[0], items[1:]
+        if label in TENSES and len(forms) == 6:
+            key = TENSE_KEYS[TENSES.index(label)]
+            if key not in indikativ:
+                indikativ[key] = dict(zip(PERSONS, forms))
+        elif label == "Infinitiv mit zu" and forms:
+            infinitiv_mit_zu = forms[0]
+        elif label == "Partizip I" and forms:
+            partizip_i = forms[0]
+        elif label == "Partizip II" and forms:
+            partizip_ii = forms[0]
+        elif len(items) == 2 and not imperative["singular"]:
+            # the Imperativ list carries no label — its two items are
+            # the singular (du) and plural (ihr) forms directly
+            imperative = {"singular": items[0], "plural": items[1]}
+    if not indikativ and not partizip_ii:
+        raise ValueError("no conjugation tables found on this page "
+                         "(is it a verb?)")
+    return {
+        "indikativ": indikativ,
+        "imperative": imperative,
+        "infinitiv_mit_zu": infinitiv_mit_zu,
+        "partizip_i": partizip_i,
+        "partizip_ii": partizip_ii,
+    }
+
+
+def parse_search(html: str) -> list[dict]:
+    """[{word, url}] — Duden headwords from the search results page."""
+    out: list[dict] = []
+    seen = set()
+    for slug in re.findall(r'href="/rechtschreibung/([^"]+)"', html):
+        if slug in seen:
+            continue
+        seen.add(slug)
+        display = slug.replace("_", " ")
+        out.append({"word": display,
+                    "url": f"{BASE}/rechtschreibung/{slug}"})
+    return out