Jimdo Docs
БесплатноНе проверенEnables semantic search over Jimdo help documentation to find relevant information and answers.
Описание
Enables semantic search over Jimdo help documentation to find relevant information and answers.
README
[!Note] Portfolio repo to apply as AI Enablement Partner
Chat with your jimdo docs.
Sample query: How to optimize SEO? Check jimdo docs
Ran qdrant-findjimdo_docs (MCP Server)
Input: { "query": "How to optimize SEO?" }
Output: {
"result": [
"Results for the query 'How to optimize SEO?'",
"
Perfect! I found relevant Jimdo Help Center articles that cover SEO basics, a practical SEO checklist, on-page tips (headings, keywords, URLs, images), and tools like rankingCoach Advance. My next step is to summarize the key, actionable recommendations and link the most relevant docs.## ✅ Quick answer — How to optimize SEO (from Jimdo docs)
Short summary: Focus on making your site discoverable and relevant: check indexation, choose target keywords, optimize titles & meta descriptions, structure content with clear headings, make URLs and images SEO‑friendly, ensure mobile responsiveness, use an SEO checklist, and consider rankingCoach Advance for guided, automated improvements.
🔧 Actionable checklist (with Jimdo doc links) Check indexation
Confirm Google has indexed your site (see: Was Google meine Jimdo Seite indexiert?). Doc: https://help.jimdo.com/hc/de/articles/115005526986-Was-ist-SEO Choose and use Keywords
Find the keywords users search with and place them in titles, headings, intro and naturally in text. Doc: https://help.jimdo.com/hc/de/articles/115005527286-Wie-optimiere-ich-meine-Texte Optimize titles & meta descriptions
Keep them descriptive and include main keyword. (See SEO checklist.) Doc: https://help.jimdo.com/hc/de/articles/115005526966-Checkliste-SEO Use meaningful headings
Short, informative headings (avoid "Startseite") using correct heading levels improves accessibility & SEO. Doc: https://help.jimdo.com/hc/de/articles/115005527286-Wie-optimiere-ich-meine-Texte Optimize URLs
Make URLs short, readable, and keyword-relevant (requires Creator Premium for manual adjustment). Doc: https://help.jimdo.com/hc/de/articles/115005527266-Wie-optimiere-ich-meine-URLs Optimize images
Use descriptive filenames, alt text, and reasonable file sizes for faster loading and accessibility. (See related items in the SEO category.) Make site mobile-friendly
Responsive sites rank better on Google — verify mobile view and responsiveness. Doc: https://help.jimdo.com/hc/de/articles/360000779286-Kann-ich-die-mobile-Ansicht-deaktivieren Submit a sitemap / use Search Console
Create and submit your sitemap to Google Search Console to help indexing. See: "Wie erstelle ich eine Sitemap?" (linked from the SEO category) Consider automated help: rankingCoach Advance
Use rankingCoach Advance for tailored, step-by-step SEO tasks and analytics. Doc: https://help.jimdo.com/hc/de/articles/115005521243-Wie-funktioniert-rankingCoach-Advance 💡 Tip: Start with the SEO Checklist (https://help.jimdo.com/hc/de/articles/115005526966-Checkliste-SEO), then implement the highest-impact items (indexation, title/meta, headings, mobile, images). Use rankingCoach Advance if you want guided, ongoing optimization.
📌 Next steps I can help with Run through the SEO checklist with you and convert items into a prioritized to‑do list. Draft optimized title/meta and heading suggestions for a specific page (share the URL or page content). Would you like me to prioritize the checklist for your site or optimize a specific page now?
Architecture
flowchart LR
subgraph AI_IDE["AI IDE tools\n(Codex, Copilot, Claude, VS Code MCP)"]
Q["Input user query"]
A["Output answers based on Jimdo docs"]
end
subgraph MCP["mcp-server-qdrant"]
R["Route query & params"]
E["Return matched docs"]
end
subgraph QD["Managed Qdrant"]
C["Jimdo docs collection (dense + sparse vectors)"]
end
subgraph Ingest["Ingestion pipeline"]
S["Sitemap crawl scrape_jimdo_docs.py"]
CSV["jimdo_docs.csv (raw pages + metadata)"]
EMB["Upload sparse and dense embeddings: generate_and_upload_embeddings.py"]
end
%% Query/answer path
Q --> R
R <--> C
C <--> E
E --> A
%% Ingestion path
S --> CSV --> EMB --> C
Prerequisites
Using the Qdrant collection via the Qdrant MCP server
For hybrid (i.e. dense+sparse) search, use the fork from Aaron Spring's branch, install via uvx: uvx --from git+https://github.com/aaronspring/mcp-server-qdrant@feat/configurable-sparse-embedding mcp-server-qdrant
VSCode via this repo
git clonethis repo- check whether
jimdo_docsappears in MCP list (see extensions) and start chat in agent mode
VSCode
- add to
.vscode/mcp.json:
{
"servers": {
"jimdo_docs": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/aaronspring/mcp-server-qdrant@feat/configurable-sparse-embedding",
"mcp-server-qdrant"
],
"env": {
"QDRANT_URL": "https://4dae4389-bdd5-43f2-8f8f-e8978a13a103.eu-central-1-0.aws.cloud.qdrant.io:6333",
"QDRANT_API_KEY": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJhY2Nlc3MiOlt7ImNvbGxlY3Rpb24iOiJqaW1kb19kb2NzX3JlY3Vyc2l2ZS03NTAtMTAwIiwiYWNjZXNzIjoiciJ9XX0.jnuAI2xvjG1sWEZtNi9MvtTUcJglupvAdcTnyk6OLx8",
"COLLECTION_NAME": "jimdo_docs_recursive-750-100",
"EMBEDDING_MODEL": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
"SPARSE_EMBEDDING_NAME": "fast-bm42-all-minilm-l6-v2-attentions",
"TOOL_FIND_DESCRIPTION": "Search Jimdo help documents using dense semantic search. Use natural language queries rather than just keywords. Returns up to 10 most relevant document excerpts. Always include document URLs in your answer to allow users to verify sources. Keep queries to qdrant very similar to the user input. Don't add too many keywords.",
"TOOL_HYBRID_FIND_DESCRIPTION": "Advanced configurable hybrid search for Jimdo help documents. Use natural language queries to leverage both semantic understanding and keyword matching. ALWAYS set these parameters: dense_limit=10 (semantic results), sparse_limit=10 (keyword results), final_limit=10 (merged results). Use fusion_method='dbsf'. Include document URLs in your answer. Keep queries to qdrant very similar to the user input. Don't add too many keywords.",
"TOOL_STORE_DESCRIPTION": "NEVER USE"
}
}
}
}
OpenCode
Create or edit your OpenCode config at one of these locations (in precedence order):
- Project-specific (highest priority):
opencode.jsonin project root - Global (user-wide):
~/.config/opencode/opencode.json
Example config:
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"jimdo_docs": {
"type": "local",
"command": [
"uvx",
"--python",
"3.12",
"--from",
"git+https://github.com/aaronspring/mcp-server-qdrant@feat/configurable-sparse-embedding",
"mcp-server-qdrant"
],
"environment": {
"QDRANT_URL": "https://4dae4389-bdd5-43f2-8f8f-e8978a13a103.eu-central-1-0.aws.cloud.qdrant.io:6333",
"QDRANT_API_KEY": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJhY2Nlc3MiOlt7ImNvbGxlY3Rpb24iOiJqaW1kb19kb2NzX3JlY3Vyc2l2ZS03NTAtMTAwIiwiYWNjZXNzIjoiciJ9XX0.jnuAI2xvjG1sWEZtNi9MvtTUcJglupvAdcTnyk6OLx8", // read only
"COLLECTION_NAME": "jimdo_docs_recursive-750-100",
"EMBEDDING_MODEL": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
"SPARSE_EMBEDDING_NAME": "fast-bm42-all-minilm-l6-v2-attentions",
"TOOL_FIND_DESCRIPTION": "Search Jimdo help documents using dense semantic search. Use natural language queries rather than just keywords. Returns up to 10 most relevant document excerpts. Always include document URLs in your answer to allow users to verify sources. Keep queries to qdrant very similar to the user input. Don't add too many keywords.",
"TOOL_HYBRID_FIND_DESCRIPTION": "Advanced configurable hybrid search for Jimdo help documents. Use natural language queries to leverage both semantic understanding and keyword matching. ALWAYS set these parameters: dense_limit=10 (semantic results), sparse_limit=10 (keyword results), final_limit=10 (merged results). Use fusion_method='dbsf'. Include document URLs in your answer. Keep queries to qdrant very similar to the user input. Don't add too many keywords.",
"TOOL_STORE_DESCRIPTION": "NEVER USE"
},
"enabled": true
}
}
}
Key points:
type: "local"— required for local MCP serverscommandmust be an array (not a string)environment(notenv) — sets env vars for the process- Restart OpenCode after config changes
- Verify with:
opencode mcp list(should showjimdo_docsas connected ✓)
Claude Code
Create or edit ~/.claude/code/config.json with the same OpenCode config structure above.
Claude Desktop
Create or edit your Claude Desktop config (e.g., on MacOS ~/Library/Application Support/Claude/claude_desktop_config.json) with the same configuration as above.
Codex CLI
Add the MCP server block to your Codex config (e.g., ~/.codex/config.toml):
[mcp_servers.jimdo_docs]
command = "uvx"
args = [
"--from",
"git+https://github.com/aaronspring/mcp-server-qdrant@feat/configurable-sparse-embedding",
"mcp-server-qdrant"
]
[mcp_servers.jimdo_docs.env]
QDRANT_URL = "https://4dae4389-bdd5-43f2-8f8f-e8978a13a103.eu-central-1-0.aws.cloud.qdrant.io:6333"
QDRANT_API_KEY = "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJhY2Nlc3MiOlt7ImNvbGxlY3Rpb24iOiJqaW1kb19kb2NzX3JlY3Vyc2l2ZS03NTAtMTAwIiwiYWNjZXNzIjoiciJ9XX0.jnuAI2xvjG1sWEZtNi9MvtTUcJglupvAdcTnyk6OLx8"
COLLECTION_NAME = "jimdo_docs_recursive-750-100"
EMBEDDING_MODEL = "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2"
SPARSE_EMBEDDING_NAME = "fast-bm42-all-minilm-l6-v2-attentions"
SPARSE_EMBEDDING_NAME = "fast-bm42-all-minilm-l6-v2-attentions"
TOOL_FIND_DESCRIPTION = "Search Jimdo help documents using dense semantic search. Use natural language queries rather than just keywords. Returns up to 10 most relevant document excerpts. Always include document URLs in your answer to allow users to verify sources. Keep queries to qdrant very similar to the user input. Don't add too many keywords."
TOOL_HYBRID_FIND_DESCRIPTION = "Advanced configurable hybrid search for Jimdo help documents. Use natural language queries to leverage both semantic understanding and keyword matching. ALWAYS set these parameters: dense_limit=10 (semantic results), sparse_limit=10 (keyword results), final_limit=10 (merged results). Use fusion_method='dbsf'. Include document URLs in your answer. Keep queries to qdrant very similar to the user input. Don't add too many keywords."
TOOL_STORE_DESCRIPTION = "NEVER USE"
- Restart Codex to pick up the MCP server.
/mcp
🔌 MCP Tools
• jimdo_docs
• Status: enabled
• Auth: Unsupported
• Command: uvx --from git+https://github.com/aaronspring/mcp-server-qdrant@feat/configurable-sparse-embedding mcp-server-
qdrant
• Env: COLLECTION_NAME=*****, EMBEDDING_MODEL=*****, QDRANT_API_KEY=*****, QDRANT_URL=*****,
SPARSE_EMBEDDING_MODEL=*****, SPARSE_EMBEDDING_NAME=*****, TOOL_FIND_DESCRIPTION=*****, TOOL_HYBRID_FIND_DESCRIPTION=*****,
TOOL_STORE_DESCRIPTION=*****
• Tools: qdrant-find, qdrant-hybrid-find, qdrant-store
• Resources: (none)
• Resource templates: (none)
Optional: Generate embeddings and upload to Qdrant
- requires
QDRANT_API_KEYwith write access (kept private)
uv run python generate_and_upload_embeddings.py --recreate \
--document-splitting recursive \
--chunk-size 750 --chunk-overlap 100
- Collections are named
jimdo_docsplus the splitting suffix (e.g.,jimdo_docs_recursive-750-100). - Use
--list-collectionsto inspect existing Qdrant collections. - Use
--n-documentsto run a quick test subset.
Document splitting
none— skip splitting; one embedding per page row (collectionjimdo_docs)recursive—RecursiveCharacterTextSplitterwithchunk_size/chunk_overlap, good default for prose (collection suffix_recursive-{chunk_size}-{chunk_overlap})tokens—TokenTextSplitter(cl100k_base) to respect token counts, useful when downstream models are token-limited (collection suffix_tokens-{chunk_size}-{chunk_overlap})--chunk-sizeand--chunk-overlapapply torecursiveandtokens; use--n-documentsfor quick dry-runs (adds_testto the collection name)- qdrant cluster UI: https://4dae4389-bdd5-43f2-8f8f-e8978a13a103.eu-central-1-0.aws.cloud.qdrant.io:6333/dashboard#/collections
Models
- Dense models for semantic search:
- paraphrase-multilingual-MiniLM-L12-v2 (default)
- jina-embeddings-v2-base-de (not working because of some onnx error)
- Sparse models for keyword search:
- Qdrant/bm25 (classic keyword search)
- Qdrant/bm42-all-minilm-l6-v2-attentions (adaptation with attention)
Установка Jimdo Docs
У этого сервера нет опубликованного пакета — он собирается из исходников. Открой репозиторий и следуй инструкции в README.
▸ github.com/aaronspring/jimdo-docs-mcpFAQ
Jimdo Docs MCP бесплатный?
Да, Jimdo Docs MCP бесплатный — установка в пару кликов через Unyly без оплаты.
Нужен ли API-ключ для Jimdo Docs?
Нет, Jimdo Docs работает без API-ключей и переменных окружения.
Jimdo Docs — hosted или self-hosted?
Доступен hosted-вариант: Unyly запускает сервер в облаке, локальная установка не обязательна.
Как установить Jimdo Docs в Claude Desktop, Claude Code или Cursor?
Открой Jimdo Docs на unyly.org, выбери вкладку своего клиента (Claude Desktop, Claude Code, Cursor) и нажми Install — конфиг сгенерируется автоматически, без правки JSON.
Похожие MCP
Notion
Read and write pages in your workspace
автор: NotionLinear
Issues, cycles, triage — from Claude
автор: LinearGoogle Drive
Search and read your Drive files
автор: Googlemindsdb/mindsdb
Connect and unify data across various platforms and databases with [MindsDB as a single MCP server](https://docs.mindsdb.com/mcp/overview).
автор: mindsdbCompare Jimdo Docs with
Не уверен что выбрать?
Найди свой стек за 60 секунд
Автор?
Embed-бейдж для README
Похожее
Все в категории productivity