Dutch Fluency

Df Check

Grade a Dutch sentence and get a direct correction with the exact rule cited + a link to the Dutch Fluency lesson that teaches it. Catches de/het article mistakes, word order, verb conjugation and auxiliary choice, spelling, prepositions, and pronouns. Text-only — works in every MCP host. (Audio version lives at dutchfluency.com.) Returns: {ok, corrected, errors[{type, span, fix, rule, lesson_url}], quick_nl, quick_en, vibe}. vibe ∈ "all_good" | "small_slip" | "big_miss".

Df Nt2 Stats

Show which KNM (Kennis Nederlandse Maatschappij) content is loaded and in which languages. Returns native_content_languages (on-disk) + available_languages (auto-translated on demand, cached) + per-language counts (chapters/facts/questions/vocab). (Other inburgering skills — Lezen/Luisteren/Spreken/Schrijven — live on dutchfluency.com; the MCP exposes KNM today.)

Df Nt2 Chapters

List the 10 KNM chapters covering Dutch society for the inburgering exam. Returns each chapter's id, title, subtitle, icon, fact_count, question_count, vocab_count. Optional lang param (nl/en/de/es/fr/tr/ar/pl/zh) — non-source languages are translated on demand and cached for 30 days. (Other inburgering skills — Lezen/Luisteren/Spreken/Schrijven — live on dutchfluency.com; the MCP exposes KNM today.)

Df Nt2 Round

Get one random KNM multiple-choice question for inburgering practice. Optional chapter_id (1–10) to constrain to one chapter. Returns question_id + question text + options (no answer leak). Answers are checked by df_nt2_check using the question_id.

Df Nt2 Check

Check an answer to a KNM question. Submit the question_id you got from df_nt2_round and the 0-based answer_idx of the option you picked. Returns whether you were correct, the expected answer text, and the explanation.

Df Nt2 Facts

Return the study-card facts for one KNM chapter. Pass chapter_id (1–10). Optional topic filter narrows by subtopic (e.g. "geography"). Each fact has id, text, topic. Chapter ids come from df_nt2_chapters.

Df Nt2 Vocab

Return KNM vocabulary for one chapter. Each entry has word_nl (the Dutch term — what the learner is studying), gloss (translation in the user's interface language), definition (longer explanation). Chapter ids come from df_nt2_chapters.

Df Certificate Issue

Mint a Dutch Fluency CEFR Certificate for a completed df_assess session. Pass session_id (from df_assess_finish) and optional user_handle for the name on the cert. Returns hash + public verification URL at www.dutchfluency.com/cert/<hash>. Unofficial credential — self-assessment anchored to CEFR descriptors, not a state exam.

Df Certificate Verify

Verify a Dutch Fluency Certificate by hash. Public — no auth required. Returns level + per-skill breakdown + issued_at + the holder's user_handle (if any) plus issued_by_authenticated_account. That flag records only that a signed-in Dutch Fluency account issued the certificate and owns the handle. It is NOT identity verification: no check ties the account holder to the name, so the name is not a confirmed identity. Returns cert_not_found if the hash is unknown or the cert was revoked.

Df Certificate Claim

Link a Dutch Fluency Certificate to an email address, no account needed. Pass the cert hash (from df_certificate_issue) and the learner's email; a confirmation link is sent to that inbox and opening it records the email as the certificate's owner. Works once per certificate. The address is stored only on the certificate record and never shown publicly. Returns ok with sent_to, or a specific error (invalid_email, cert_not_found_or_revoked, already_claimed, too_many_requests, email_send_failed).

Df Progress List

List YOUR OWN Dutch Fluency assessment history. Requires a signed-in account: send the learner's bearer token. Returns up to 20 most-recent certificates issued by that account, newest-first. user_handle is an optional filter within your own certificates, not a way to look up another learner. Each entry includes hash, global_cefr, per-skill levels, items_served, issued_at, and the public verification URL. To check someone else's certificate, use df_certificate_verify with the hash they shared.

Df Progress Compare

Compare two Dutch Fluency Certificates. Pass two distinct cert hashes — returns per-skill CEFR deltas (up/down/flat), days_between, and the earlier + later level. Useful for showing improvement between retakes.

Df Assess Start

Start a new Dutch CEFR placement session. Returns session_id + the first framing item (motivation question). After 3 framing items, the flow moves into ~12 adaptive reading + grammar items, then a level-matched writing prompt. Optional branding="neutral" disables the DFL personality copy and returns a clean third-party-friendly response. Audio-based skills (listening, pronunciation, speaking) live on dutchfluency.com — the MCP protocol only carries text.

Df Assess Answer

Submit the picked option for the current placement test item. Pass session_id (from df_assess_start), the item_id you just received, and picked (the option id A/B/C/D). Returns the next item or — after 12 receptive items are done — the writing prompt with phase="writing".

Df Assess Submit Writing

Submit the learner's Dutch writing sample for the placement test. Pass session_id (must have completed the receptive phase) and text (≥20 chars). Returns CEFR level, per-rubric scores (task completion, grammar, vocab, coherence, register), strengths, errors with corrections + lesson URLs, and next-focus skills. The final report is computed by df_assess_finish.

Df Assess Finish

Compute the final placement report. Pass session_id. Returns per-skill CEFR levels (lezen, grammatica, schrijven), global CEFR, can-do statements at that level, weakest-skill focus, missed teaching tags, and a 30-60-90 day learning path. A shareable certificate can be minted from the same session_id with df_certificate_issue.

Df Assess Grade Sample

Single-shot CEFR estimate for any Dutch text. No session required. Paste 20+ characters and get back a CEFR level (A1–C1), per-rubric scores, strengths, errors with corrections + lesson URLs, and next-focus skills. Useful for ad-hoc level-checking; certificates come from the full df_assess_start → df_assess_finish flow.

Df Speech Analyse

Analyse spoken Dutch from a lesson transcript or pasted Dutch text. Accepts WEBVTT, "Name: text" turns, or prose. Splits speakers, counts Dutch vs English, floor-holders (yeah/okay/I think), and repeated grammar cues. If the host is Dutch-heavy and the learner is English-heavy in the same conversation, flags likely_asr_english (Zoom / English-locked ASR on accented Dutch) so you do not coach "drop the English" on a Dutch lesson. Does not score pronunciation. Not a CEFR certificate.