Skip to main content
Glama
modulrdesign

Thermomix-MCP

by modulrdesign
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Hybrid

Thermomix-MCP

Rezepte aus beliebigen Quellen (Rezept-Websites, YouTube, Instagram, TikTok, KI) in ein Thermomix-Rezept verwandeln — und wahlweise direkt in Cookidoo → „Meine Kreationen" hochladen, inkl. Guided-Cooking-Schritten und Bild, sodass es am TM7 abrufbar ist.

Zwei Betriebsarten:

  • MCP-Server (Chat-Flow) — Claude bekommt die Rohdaten und baut das Rezept.

  • Webhook (headless) — nimmt eine geteilte URL entgegen und macht die Umwandlung selbst per Anthropic-API. Ideal für einen „aus jeder App teilen"-Flow.

Kein offizielles Vorwerk-/Cookidoo-API. Der Cookidoo-Upload nutzt eine reverse-engineerte Schnittstelle (cookidoo-api

  • rohe created-recipes-Endpoints) und ein aktives Cookidoo-Abo. Das kann jederzeit brechen. Der TM7 selbst lässt sich nicht fernsteuern — nur Content-Sync über den Account.

Quellen

Quelle

Weg

Rezept-Websites (Chefkoch & Co.)

schema.org/Recipe / JSON-LD (recipe-scrapers)

YouTube

Beschreibung + Transkript

Instagram / TikTok

Caption (oEmbed), optional Tonspur-Transkription (Whisper)

KI-generiert

Claude erstellt das Rezept ohne Quelle

Enthält ein Post kein Rezept (z. B. TikTok-Caption „recipe on the blog"), sucht der Deep-Hunt die Original-Rezeptseite per Websuche und transkribiert notfalls die Tonspur.

Related MCP server: yes_chef_mcp

MCP-Tools & Prompts

Tool

Zweck

fetch_recipe(url)

Rezept-Rohdaten aus Web/YouTube/Instagram/TikTok

find_recipe_image(query)

Fallback-Bildsuche (Pexels)

save_thermomix_recipe(markdown, filename)

Rezept-Karte als .md speichern

upload_thermomix_recipe(recipe, image_url=…)

Rezept nach Cookidoo hochladen (Guided)

Prompts: thermomix_guide (TM-Konvertierungsregeln), cookidoo_guide (Modus-Zuordnung).

Setup

uv sync                         # Kern
uv sync --extra cookidoo        # + Cookidoo-Upload
uv sync --extra whisper         # + Audio-Transkription (Social-Reels)
cp .env.example .env            # Keys/Optionen eintragen

ffmpeg wird für YouTube-/Social-Audio gebraucht (brew install ffmpeg).

In Claude Code einbinden

claude mcp add thermomix -- uv --directory /pfad/zu/thermomix-mcp run thermomix-mcp

Dann z. B.: „Mach mir aus dieser Chefkoch-URL ein Thermomix-Rezept."

Webhook — „aus jeder App teilen"

uv run thermomix-webhook        # startet auf 0.0.0.0:8787

Endpunkte:

  • GET /health → Status

  • POST /recipe mit {"url": "…", "caption": "…"(optional)} und Header X-Auth-Token: <WEBHOOK_TOKEN> → extrahiert, wandelt um, speichert die Karte, lädt (wenn konfiguriert) nach Cookidoo hoch. Rezeptlose Video-Posts → sofort 202 + Deep-Hunt im Hintergrund.

Für den Zugriff von unterwegs den Webhook über einen Tunnel freigeben (z. B. Tailscale Funnel oder Cloudflare Tunnel — feste HTTPS-URL, keine Portfreigabe). Ein iOS-Kurzbefehl im Teilen-Menü POSTet die geteilte URL dorthin. Vor jeder öffentlichen Exposure WEBHOOK_TOKEN setzen.

Konfiguration (.env)

Variable

Zweck

THERMOMIX_OUTPUT_DIR

Zielordner für Rezept-Karten (Default ./recipes)

PEXELS_API_KEY

Fallback-Bildsuche (kostenlos)

WHISPER_MODEL

tiny/base/small/medium (Default base)

YTDLP_COOKIES_FROM_BROWSER

safari/chrome/… für Login-Wände bei Instagram/TikTok

ANTHROPIC_API_KEY

Webhook: für die headless-Umwandlung nötig

THERMOMIX_LLM_MODEL

Webhook-Modell (Default Haiku; claude-sonnet-5 für mehr Qualität)

WEBHOOK_TOKEN

Webhook: Auth-Token — vor öffentlicher Exposure zwingend

WEBHOOK_HOST / WEBHOOK_PORT

Webhook-Bind (Default 0.0.0.0:8787)

COOKIDOO_EMAIL / COOKIDOO_PASSWORD

Cookidoo-Upload (braucht aktives Abo)

Lizenz

MIT

Available Tools

4 tools
fetch_recipeA

Liest ein Rezept aus einer beliebigen Quelle und gibt Rohdaten zurück.

Unterstützt: Rezept-Websites (Chefkoch & Co. via schema.org/Recipe), YouTube (Beschreibung + Transkript), Instagram & TikTok (Caption + optional Audio-Transkript).

Nächster Schritt für dich (das aufrufende Modell): Wandle diese Rohdaten in ein Thermomix-Rezept um — folge den Regeln aus dem Prompt thermomix_guide. Nimm das image_url als Rezeptbild; fehlt es, hol per find_recipe_image einen Fallback. Speichere am Ende mit save_thermomix_recipe. Beachte das Feld notes.

ParametersJSON Schema
NameRequiredDescriptionDefault
urlYes

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A3.9/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden. It discloses that the tool returns raw data and mentions supported sources, but it does not discuss auth, rate limits, or error behavior. It adequately suggests a non-destructive read operation.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is moderately concise, with a clear structure separating supported sources from post-fetch instructions. Every sentence adds value, though the extra guidance for the calling model could be considered slightly beyond the tool's core description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (multiple source types) and the existence of an output schema, the description is fairly complete. It covers input, output, and follow-up actions, though it omits error specifics.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With schema description coverage at 0%, the description adds meaning by implying that the 'url' parameter should point to a supported recipe source. However, it lacks details like format examples or validation rules.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the verb ('liest ein Rezept') and resource ('aus einer beliebigen Quelle') and specifies supported sources, distinguishing it from sibling tools like find_recipe_image and save_thermomix_recipe.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explains when to use this tool (to fetch raw recipe data from a URL) and provides a clear next-step pipeline (convert to Thermomix recipe, save). It lacks explicit when-not-to-use or alternatives, but the context is clear.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

find_recipe_imageA

Sucht ein passendes Gerichtsfoto (Fallback, wenn das Rezept kein eigenes Bild hat).

Bevorzuge immer das image_url aus fetch_recipe. Kurzer, konkreter Suchbegriff (z.B. "Kürbissuppe Teller"). Braucht einen PEXELS_API_KEY in der .env.

ParametersJSON Schema
NameRequiredDescriptionDefault
queryYes

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A4.6/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

The description discloses that it requires a PEXELS_API_KEY in the .env file, which is important for the agent to know. However, it does not specify rate limits, error behavior, or details about the external API call. With no annotations, the description carries the full burden; the key disclosure adds value.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is three sentences, efficiently front-loading the purpose and then providing usage guidance and parameter advice. No unnecessary words.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a simple tool with one parameter and an existing output schema, the description covers purpose, usage guidelines, parameter semantics, and an environment requirement. It is sufficiently complete for an agent to use correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema has a single string parameter 'query' with no description (0% coverage). The description compensates by advising a 'short, concrete search term' with an example, adding meaning beyond the schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose: searching for a matching dish photo as a fallback when the recipe has no own image. It distinguishes itself from siblings like fetch_recipe by positioning this as a fallback.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicit guidance is given: first prefer the image_url from fetch_recipe, and only use this tool as fallback. Also provides an example of a concrete search term.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

save_thermomix_recipeA

Speichert die fertige Thermomix-Rezept-Karte als Markdown-Datei.

markdown = vollständige Karte (mit YAML-Frontmatter, wie im Prompt thermomix_guide beschrieben). filename optional — sonst aus der Überschrift abgeleitet. Zielordner: THERMOMIX_OUTPUT_DIR oder /recipes.

ParametersJSON Schema
NameRequiredDescriptionDefault
filenameNo
markdownYes

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A3.9/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description bears full responsibility. It mentions the output format and target folder but lacks details on overwrite behavior, error handling, or prerequisites. The basic file-saving operation is described, but depth is limited.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is concise, using two sentences to convey the core purpose and parameter details. It front-loads the main action. However, it could be slightly more structured or include a brief note on behavior.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's low complexity (2 parameters) and presence of an output schema, the description covers essential aspects: purpose, parameters (meaning), and target folder. It lacks alternative tool guidance but is otherwise adequate for a simple save operation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With 0% schema description coverage, the description effectively compensates by explaining what `markdown` contains (full card with YAML frontmatter) and that `filename` is optional, derived from the heading. It adds meaningful context beyond bare parameter names.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly specifies the tool's action: saving a finished recipe card as a Markdown file. It distinguishes from siblings (fetch, find, upload) by focusing on local file saving.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explains how to use the parameters but does not provide explicit guidance on when to use this tool versus alternatives like upload_thermomix_recipe. Usage context is implied but not elaborated.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

upload_thermomix_recipeA

Lädt ein Rezept in Cookidoo „Meine Kreationen" hoch → Guided Cooking am TM7.

recipe = strukturiertes JSON wie im Prompt cookidoo_guide beschrieben (name, ingredients, steps[{text, mode?}], totalTime, prepTime, yield, tools). Setze mode je Schritt nur, wo die Werte exakt in einen Cookidoo-Modus passen (Details im Prompt) — sonst bleibt der Schritt Text. image_url = Rezeptfoto (nimm das image_url aus fetch_recipe); wird zu Cloudinary hochgeladen und erscheint in der Cookidoo-App. Braucht COOKIDOO_EMAIL/COOKIDOO_PASSWORD in der .env + aktives Cookidoo-Abo. Antwort enthält recipe_id und guided_buttons (Anzahl gesetzter Guided-Modi).

ParametersJSON Schema
NameRequiredDescriptionDefault
recipeYes
image_urlNo

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A3.8/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries full burden. It discloses that the tool uploads to Cookidoo, handles image upload to Cloudinary, requires credentials and subscription, and returns recipe_id and guided_buttons. This provides good behavioral insight beyond the input schema.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is front-loaded with the main purpose and then provides necessary details. It is concise for the complexity, with each sentence adding value (structure of recipe, mode guidelines, image handling, requirements). No redundancies.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given two parameters, an output schema (exists), and sibling tools, the description covers what the tool does, how to use the parameters, prerequisites (env vars, subscription), and response fields. It lacks explicit differentiation from siblings but is otherwise complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 0%, so the description must add meaning. It does so by explaining the 'recipe' parameter as structured JSON (name, ingredients, steps with optional mode, etc.) and 'image_url' as a photo from fetch_recipe. This adds significant context beyond the generic object and string schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool uploads a recipe to Cookidoo 'Meine Kreationen' and enables Guided Cooking on TM7. The verb 'Lädt... hoch' is specific, and the resource is identified. However, it does not explicitly differentiate from the sibling 'save_thermomix_recipe', which could be similar.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides usage context: requires COOKIDOO_EMAIL/COOKIDOO_PASSWORD in .env and an active subscription. It implies when to use (uploading structured recipes), but does not give explicit when-not or alternatives (e.g., vs save_thermomix_recipe).

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 4 tool updatesv0.1.0
    • First observedfetch_recipe
    • First observedfind_recipe_image
    • First observedsave_thermomix_recipe
    • First observedupload_thermomix_recipe
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A4.1/5.0

Scored across 4 tools

Disambiguation5/5

Each tool has a distinct and non-overlapping purpose: fetching raw recipe data, finding a fallback image, saving as markdown, and uploading to Cookidoo. No ambiguity.

Naming Consistency5/5

All tool names follow a consistent verb_noun pattern with underscores (e.g., fetch_recipe, find_recipe_image), making them predictable and easy to understand.

Tool Count4/5

4 tools is on the lower side but appropriate for the focused domain of Thermomix recipe processing. Each tool serves a clear step in the workflow without unnecessary bloat.

Completeness4/5

The tool set covers the core workflow: fetching, image fallback, saving locally, and uploading. The conversion step is handled by the model via a prompt, so no gap exists.

Maintenance

ActivityStale
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • yes_chef_mcp

    ltmorro
    F
    license
    Not graded
    quality
    D
    maintenance
    Enables AI assistants to search recipes, compose nutritionally balanced meals, optimize weekly meal plans based on macro targets for family members, and generate consolidated grocery lists from a personal recipe database.
    -

  • cookidoo-mcp

    sisques-labs
    A
    license
    Not graded
    quality
    B
    maintenance
    MCP server for Cookidoo, enabling AI tools to search recipes, manage shopping lists, and retrieve account and subscription information.
    2
    GPL 3.0

  • cookidoo-mcp

    jsaedtler
    A
    license
    A
    quality
    B
    maintenance
    Enables LLMs to upload and manage Thermomix recipes on Cookidoo, including structured steps with Thermomix settings, images, and access to shopping lists and meal planning.
    11
    MIT