Lingard MCP server

• Registry nameapp.lingard/vocabverified in the MCP registry

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

• Stateanswerschecked 2026-10-30
• Tools15
• Server calls itselflingard-vocab
• Server version0.5.9
• Protocol version2025-06-18

On 2026-10-30 the server answered initialize and tools/list without a sign-in. First answer in 113 ms.

• Its tools do save new words and review them with spaced repetition, but not read graded books and articles at their own level.
• 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.

• INITE MCP scoreafter a weeka score needs 7 daily checks: 5 so far
• What an agent can do with it1 of 2 use cases
• Answeredmeasuring5 of the 7 daily checks it takes to print a share
• Median first answer134 ms
• Space in an agent's context≈ 10,879 tokensthe definitions of all its tools (15), which an agent carries on every turn
• Tool definitionsB · 87 of 100
• Typed results15 of 15 toolsdeclared outputSchema: results come typed, read once

• Parameters without a description: in 12 of 15

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 →

• search

  Searches the user's live Lingard account: vocabulary, learning context, English profile, totals, collection titles, saved words, flashcards, and missing vocabulary fields. Results are read from Firestore at the time of the call.

• fetch

  Fetches the full live Lingard JSON record for a search-result id (e.g. "learning-context:en") or a Lingard resource URI such as lingard://learning-context/en.

• lingard_get_learning_context

  Returns the user's current Lingard learning context, read live from their account: English profile, vocabulary totals, saved collections and their titles, card counts, and missing-field stats. Earlier answers in the conversation may be…

• lingard_get_vocab_collection

  Returns live cards from one Lingard vocabulary collection by UUID or title, including curated Lingard collections and the "Words not in a collection" bucket of cards saved while reading. With includeExamples:true it also returns each…

• lingard_get_collection_gaps

  Returns live Lingard cards in a collection that are missing requested fields such as translation, POS, pronunciation, examples, or image. A curated Lingard collection always reports zero gaps: its words come from the Lingard catalog and…

• lingard_get_word_suggestion_context

  Returns live Lingard profile level, topic, collection context, and the words the user already has: the context for suggesting level-appropriate new cards that are not duplicates.

• lingard_preview_vocab_changes

  Previews a Lingard vocabulary collection/card change without writing to Firestore. Cards are checked as on commit. Unless allowIncompleteCards is true, a new card needs title, translate, pos, pron/ipa, syns, and examples with…

• lingard_commit_vocab_changes

  Commits a previously previewed or freshly supplied Lingard vocabulary change. Unless allowIncompleteCards is true, a new card needs title, translate, pos, pron/ipa, syns, and examples with translations; image is optional. A German or…

• lingard_create_vocab_collection

  Creates a Lingard vocabulary collection, optionally with initial cards; commit=true writes it. Unless allowIncompleteCards is true, a new card needs title, translate, pos, pron/ipa, syns, and examples with translations; image is optional…

• lingard_add_words_to_collection

  Adds cards to an existing Lingard collection, deduplicating by normalized word title; commit=true writes them. Only collections with writable="full" accept new cards: curated Lingard collections are read-only and the unfiled bucket takes…

• lingard_fill_missing_card_fields

  Fills empty fields on existing Lingard cards. It never overwrites or clears a field that already has a value. Works on the user's own collections and on the "Words not in a collection" bucket; curated Lingard collections are read-only…

• lingard_update_example_translations

  Rewrites only the translation line of examples that already exist on Lingard cards; the example sentence, its {} highlights and its audio/typing markers are left exactly as they are. It changes, adds, or removes example translations; an…

• lingard_attach_card_images

  Attaches supplied images to existing Lingard cards. It does not generate images: each entry carries the image itself as imageBase64 or dataUri, or an imageUrl that the server downloads. PNG, JPEG, WebP, and SVG inputs are converted to…

• lingard_get_learning_pauses

  Lists, for one language, the user's collections and sources (books, the news) and which of them are on pause. A collection or source on pause gives the automatic «Now learning» sessions no words — no new words and no reviews — and is not…

• lingard_set_learning_pauses

  Puts collections and sources on pause or takes them off it, for one language — e.g. "study only my Travel collection this week", "pause the news words", "resume everything". action: "pause" or "resume" the listed ones; "only" — resume the…