Islam West Africa Collection (IWAC)

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Read-only access to the Islam West Africa Collection via Hugging Face datasets.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "Islam West Africa Collection (IWAC)": {
      "url": "https://islam.zmo.de/mcp/",
      "transport": "streamable-http"
    }
  }
}

Measured tools (39)

• search read-only

  Search the Islam West Africa Collection across newspaper articles, Islamic publications, archival documents, academic references, audiovisual recordings, photographs, and the authority index (persons/places/organisations/events/subjects). Pass ONE concept or name — e.g. 'Tijaniyya', 'laïcité', 'Sheikh Gumi', 'pèlerinage'. Matching is accent- and case-insensitive; a multi-word query requires every word to appear somewhere in the item, so prefer a single concept per call. Write query strings and concept keywords in French for press/publication/document/index discovery even when the user's report language is not French. Academic references are multilingual, so try French and English title/abstract terms when relevant; metadata/filter labels remain French. Use the French transliteration of Islamic terms (Tabaski not 'Eid al-Adha', charia not 'sharia', Maouloud not 'Mawlid'). Returns {results:[{id,title,url,category}], ranking}; each result's `category` names its subset and the `ranking` field documents the ordering. Pass an id to `fetch` to read the full text. For filtered queries (by country, date, or newspaper) use the search_* tools instead.

  Parameters:

  • query (required) — string — One concept, name, or short phrase; use French concept terms for primary sources, and French/English terms for references
  • limit — integer — Max results across all categories. Default 20, max 50.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• fetch read-only

  Retrieve the full text and metadata of one IWAC item by an id returned from `search` (format '<category>:<number>', e.g. 'articles:28576'). Returns {id, title, text, url, metadata}: `text` is the item's OCR / abstract / transcription / description, `url` is the canonical islam.zmo.de link to cite, and `metadata` holds the remaining fields (author, date, country, newspaper, AI sentiment, …). Categories: articles, publications, references, documents, index, audiovisual, images.

  Parameters:

  • id (required) — string — Item id from search, e.g. 'articles:28576' or 'references:11045'

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• explore_corpus read-only

  Selection workbench: items, concordance, coverage, compare (requires comparison), attention, aliases, or paginated manifest/CSL-JSON/BibTeX exports. Exact filters intersect. Coverage measures archive availability. Exports identify this page and dataset snapshot; compare snapshot IDs across pages. Schema discovery: iwac://datasets/{subset}.

  Parameters:

  • mode — string
  • limit — integer
  • offset — integer
  • subset — string
  • selection — object
  • comparison — object

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_articles read-only

  Search IWAC newspaper articles by keyword (title + OCR + AI abstracts, French and English), country, newspaper, subject, and date range. Use French concept keywords regardless of the user's report language. Matching is accent- and case-insensitive.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • limit — integer — Default 20, max 100 (10 and 25 with with_description)
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Togo (accents optional)
  • date_to — string — YYYY-MM-DD (or YYYY)
  • keyword — string — Concept keyword; substring match on title, OCR text, and the French and English AI abstracts. Prefer French for the OCR; an English term still matches via the English abstract
  • subject — string
  • date_from — string — YYYY-MM-DD (or YYYY)
  • newspaper — string
  • hijri_year — integer — Islamic (Umm al-Qura) year, e.g. 1445
  • hijri_month — string — Islamic lunar month: 1-12, or a name (Ramadan, Chaabane, Chawwal, Dhu al-Hijja). Pulls the articles behind an observance peak — matches only items with a full YYYY-MM-DD date.
  • keyword_mode — string
  • keyword_aliases — array
  • with_description — boolean — Include each article's ~500-char AI abstract (description_ai) for triage without get_article. Adds ~125 tokens/row, so `limit` defaults to 10 and caps at 25 while this is on.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_article read-only

  Get one article (by id): full metadata, the AI abstract (description_ai), AI sentiment, and OCR text, in 25k-char parts (follow `next_offset`). Pass a `keyword` to get ~2000-char excerpts around each match instead.

  Parameters:

  • article_id (required) — integer
  • offset — integer — Read from this character on: the previous part's next_offset
  • keyword — string — Return excerpts around matches instead of the full OCR (accent-insensitive)
  • max_excerpts — integer — Default 10, max 25
  • context_chars — integer — Default 2000, max 5000

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• semantic_search_articles read-only

  Semantic search over article title + full text, including articles whose OCR is not public. Query in any language. Runs the corpus model, EmbeddingGemma 2, locally with no API key (first use downloads ~350 MB). similarity_score only ranks within one query.

  Parameters:

  • query (required) — string — Natural-language query, any language
  • limit — integer — Default 10, max 50
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Togo (accents optional)
  • date_to — string — YYYY-MM-DD (or YYYY)
  • date_from — string — YYYY-MM-DD (or YYYY)
  • newspaper — string

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_by_sentiment read-only

  Filter articles by gpt-5-6-luna sentiment labels (accent/case-insensitive exact match). One model's reading, not a consensus — 4 other models scored the same articles and often disagree; get_sentiment_distribution with model:"all" shows by how much. `subjectivity` is much the weakest of the three scales, so treat a set selected on it as a lead to read rather than as a finding.

  Parameters:

  • limit — integer — Default 20, max 100
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Togo (accents optional)
  • subject — string
  • disputed — string — polarite | centralite | subjectivite — keep only articles the panel SPLIT on for that field (French field names, as stored). Selects contested readings, not a sentiment value.
  • polarity — string — Très positif | Positif | Neutre | Négatif | Très négatif | Non applicable
  • centrality — string — Très central | Central | Secondaire | Marginal | Non abordé
  • subjectivity — string — Très objectif | Plutôt objectif | Mixte | Plutôt subjectif | Très subjectif — least to most subjective. Unscored where the model answered Non abordé, so this filter also excludes those.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_sentiment_distribution read-only

  Aggregate AI labels over a shared selection. model:"all" compares independent annotators and their coverage; model:"consensus" reads the stored panel majority. Choose compare_models and agreement_field for pairwise Cohen's kappa and quadratic weighted kappa, with explicit denominators. Labels are interpretations, not ground truth; subjectivity is weak evidence and its caveat must accompany findings.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • model — string — gpt-5-6-luna | mistral-small-2603 | deepseek-v4-flash-0731 | gemma-4-31b-it | qwen3-8-27b | all | consensus; default gpt-5-6-luna. Vendor aliases accepted; retired models rejected.
  • country — string — Exact country name (accents optional)
  • date_to — string
  • keyword — string
  • subject — string — Exact tag
  • date_from — string — Inclusive date bounds: YYYY[-MM[-DD]]
  • newspaper — string — Newspaper/periodical substring
  • hijri_year — integer
  • hijri_month — string
  • keyword_mode — string
  • compare_models — array — Pair of model IDs or aliases; requires model:"all"
  • agreement_field — string
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_index read-only

  Search the IWAC authority index (persons, places, organisations, events, subjects) by name. Accent/case-insensitive.

  Parameters:

  • keyword (required) — string — Search term matched against the entry title
  • limit — integer — Default 20, max 100
  • offset — integer
  • index_type — string — Exact type (accents optional), validated against: Personnes | Lieux | Organisations | Événements | Sujets | Notices d'autorité. An unrecognised value returns an error listing the valid types.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_index_entry read-only

  Get full details of an index entry by id (raw dataset columns, French names — Titre, Prénom, Coordonnées…).

  Parameters:

  • entry_id (required) — integer

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• list_subjects read-only

  List sujets from the IWAC index, sorted by frequency (most-referenced first).

  Parameters:

  • limit — integer — Default 50, max 200
  • offset — integer

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• list_locations read-only

  List lieux from the IWAC index, sorted by frequency (most-referenced first). The optional 'country' filter selects entries that APPEAR IN records from that country (mentioned-in, not located-in), ranked by collection-wide 'frequency' — so foreign and cross-border entries can appear. Nigeria returns none here (index frequency is computed from articles + publications + references, which have no Nigerian items — Nigeria is audiovisual only).

  Parameters:

  • limit — integer — Default 50, max 200
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional). Selects lieux MENTIONED IN records from that country, not entities located there

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• list_persons read-only

  List personnes from the IWAC index, sorted by frequency (most-referenced first). The optional 'country' filter selects entries that APPEAR IN records from that country (mentioned-in, not located-in), ranked by collection-wide 'frequency' — so foreign and cross-border entries can appear. Nigeria returns none here (index frequency is computed from articles + publications + references, which have no Nigerian items — Nigeria is audiovisual only).

  Parameters:

  • limit — integer — Default 50, max 200
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional). Selects personnes MENTIONED IN records from that country, not entities located there

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_collection_stats read-only

  Overall statistics for every IWAC subset, including `fulltext_coverage` — how many items in each subset actually carry searchable full text in this dataset. Read that before treating any keyword count as a full-text census.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_newspaper_stats read-only

  Per-newspaper article counts and date ranges.

  Parameters:

  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Togo (accents optional)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_country_comparison read-only

  Compare article counts, newspaper counts, date ranges, and gpt-5-6-luna polarity across countries.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_temporal_distribution read-only

  Counts by Gregorian or Hijri year/month, or pooled lunar month. Group by country/newspaper. Partial dates and missing lunar dates are disclosed. normalize_by returns numerator counts and per-bucket denominators in the same country/outlet/date scope, removing thematic filters. Shares describe IWAC holdings.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • subset — string — articles (default) | publications | references | documents | audiovisual | images
  • country — string — Exact country name (accents optional)
  • date_to — string
  • keyword — string
  • subject — string — Exact tag
  • calendar — string — gregorian (default) | hijri — bucket by the Islamic (Umm al-Qura) calendar
  • group_by — string — country | newspaper — one distribution per group value
  • date_from — string — Inclusive date bounds: YYYY[-MM[-DD]]
  • newspaper — string — Newspaper/periodical substring
  • hijri_year — integer
  • granularity — string — year (default) | month | lunar_month (all years collapsed into 12 lunar months; needs calendar=hijri)
  • hijri_month — string
  • keyword_mode — string
  • normalize_by — string — Share per bucket in the same country/outlet/date scope; searchable restricts both counts to full text. Thematic filters are removed from the denominator.
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_topic_distribution read-only

  Counts of offline LDA topic assignments, optionally over time. Topic labels are interpretive model outputs; exact topic_id filters select assignments rather than keyword matches. Reports unclassified coverage.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • top_n — integer — Topics given their own band in over_time (default 8, max 15)
  • subset — string — articles (default) | references
  • country — string — Exact country name (accents optional)
  • date_to — string
  • keyword — string
  • subject — string — Exact tag
  • min_prob — number — 0-1; keep only assignments at or above this probability
  • date_from — string — Inclusive date bounds: YYYY[-MM[-DD]]
  • newspaper — string — Newspaper/periodical substring
  • over_time — boolean — Also return per-year counts for the leading topics
  • hijri_year — integer
  • hijri_month — string
  • keyword_mode — string
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_field_distribution read-only

  Rank exact metadata values, splitting pipe-separated tags. Reports missing metadata and optional coverage over time. Choose spatial for mentioned places, country for stored country tags, or subject/author/language.

  Parameters:

  • field (required) — string — subject | spatial | author | language | newspaper | country
  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • top_n — integer — Values returned (default 25, max 100)
  • subset — string — articles (default) | publications | references
  • country — string — Exact country name (accents optional)
  • date_to — string
  • keyword — string
  • subject — string — Exact tag
  • date_from — string — Inclusive date bounds: YYYY[-MM[-DD]]
  • newspaper — string — Newspaper/periodical substring
  • over_time — boolean — Also return the per-year share of items carrying a value
  • hijri_year — integer
  • hijri_month — string
  • keyword_mode — string
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_cooccurrence read-only

  Count pairs of metadata values co-tagged on the same items. Matrix/network views show descriptive associations, not causal links. Filters apply before pair counting; top_n caps the vocabulary.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • field — string — subject (default) | spatial | author | language
  • top_n — integer — Values on each axis (default 15, max 30)
  • subset — string — articles (default) | publications | references
  • country — string — Exact country name (accents optional)
  • date_to — string
  • keyword — string
  • subject — string — Exact tag
  • date_from — string — Inclusive date bounds: YYYY[-MM[-DD]]
  • newspaper — string — Newspaper/periodical substring
  • hijri_year — integer
  • hijri_month — string
  • keyword_mode — string
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_place_distribution read-only

  Place tags joined to authority coordinates. Separates publication countries from mentioned places; missing geocodes are counted explicitly.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • top_n — integer — Geocoded places returned (default 60, max 200)
  • subset — string — articles (default) | publications | references
  • country — string — Exact country name (accents optional)
  • date_to — string
  • keyword — string
  • subject — string — Exact tag
  • date_from — string — Inclusive date bounds: YYYY[-MM[-DD]]
  • newspaper — string — Newspaper/periodical substring
  • hijri_year — integer
  • hijri_month — string
  • keyword_mode — string
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_semantic_map read-only

  Stable-hash sample of matching stored embeddings, projected with PCA. Reports eligible, invalid, missing and capped counts, seed and explained variance. Group counts describe the sample. Coordinates are chart-only; this is not the published UMAP landscape. Needs no API key.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • limit — integer — Items projected (default 300, max 2000)
  • subset — string — articles (default) | documents | publications | references
  • country — string — Exact country name (accents optional)
  • date_to — string
  • keyword — string
  • subject — string — Exact tag
  • color_by — string — country | newspaper | subject | lda_topic_label | polarity (gpt-5-6-luna's label)
  • date_from — string — Inclusive date bounds: YYYY[-MM[-DD]]
  • newspaper — string — Newspaper/periodical substring
  • hijri_year — integer
  • hijri_month — string
  • keyword_mode — string
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_similar_items read-only

  Nearest stored embeddings by cosine similarity. Articles and documents lead with the dataset's published top 10. Scores run high (unrelated ~0.7) and rank reading candidates; >= 0.97 suggests a near-duplicate, but no threshold proves a reprint. Returns source IDs and dates. Corpus-wide candidate generation belongs offline.

  Parameters:

  • id (required) — string — Item id — either a bare o:id ('3064') or the namespaced form search returns ('articles:3064')
  • limit — integer — Neighbours returned (default 12, max 50)
  • subset — string — articles (default) | documents | publications | references
  • min_score — number — Minimum cosine similarity (-1 to 1)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_lexical_metrics read-only

  Text length, lexical richness and readability by country/newspaper/year. Metrics are precomputed; readability is French-specific and excludes other languages. Missing values are disclosed. Compare distributions as descriptive evidence.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • top_n — integer — Groups returned when grouping by newspaper (default 20, max 60)
  • country — string — Exact country name (accents optional)
  • date_to — string
  • keyword — string
  • subject — string — Exact tag
  • group_by — string — year (default) | newspaper | country
  • date_from — string — Inclusive date bounds: YYYY[-MM[-DD]]
  • newspaper — string — Newspaper/periodical substring
  • hijri_year — integer
  • hijri_month — string
  • keyword_mode — string
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_publications read-only

  Search Islamic publications (periodical issues, books). `keyword` matches title, subject, table of contents, and full OCR text (TOC hits come back as matching_toc_entries); use French concept keywords regardless of the user's report language. Filter by newspaper/series, subject, country and year. Use list_periodicals to discover series titles, and get_publication_fulltext for keyword excerpts from a single issue.

  Parameters:

  • exact — object — AND values. Fields: subject/spatial/author/language/country/newspaper/topic_id/min_prob/scored_by; country_raw=whole cell; polarity:<model>.
  • limit — integer — Default 20, max 100
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Togo (accents optional)
  • date_to — string — Latest year, YYYY
  • keyword — string — French concept keyword; substring match on title + subject + table of contents + OCR (accent-insensitive)
  • subject — string — Subject tag (~87% of issues are tagged)
  • date_from — string — Earliest year, YYYY
  • newspaper — string — Periodical/series title (see list_periodicals)
  • hijri_year — integer — Islamic (Umm al-Qura) year, e.g. 1445
  • hijri_month — string — Islamic lunar month: 1-12, or a name (Ramadan, Chaabane, Chawwal, Dhu al-Hijja). Matches only issues with a full YYYY-MM-DD date — ~83% of them.
  • keyword_mode — string
  • keyword_aliases — array

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• list_periodicals read-only

  List the Islamic periodical/series titles in the publications subset, with issue counts and year ranges. Use the returned newspaper value as the `newspaper` filter on search_publications.

  Parameters:

  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Togo (accents optional)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_publication_fulltext read-only

  Full OCR text of a publication in 25k-char parts (follow `next_offset`), or ~2000-char excerpts around keyword matches (accent-insensitive; capped: see match_count vs excerpts_returned).

  Parameters:

  • publication_id (required) — integer
  • offset — integer — Read from this character on: the previous part's next_offset
  • keyword — string
  • max_excerpts — integer — Default 10, max 25
  • context_chars — integer — Default 2000, max 5000

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• semantic_search_publications read-only

  Semantic search over publication tables of contents, present for ~40% of issues (none for An-Nasr Vendredi, 130 of 695 Islam Info issues): use search_publications for the rest. Rows show the first 600 characters of each table. Query in any language. Runs the corpus model, EmbeddingGemma 2, locally with no API key (first use downloads ~350 MB). similarity_score only ranks within one query.

  Parameters:

  • query (required) — string — Natural-language query, any language
  • limit — integer — Default 10, max 50
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Togo (accents optional)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_references read-only

  Search academic references (journal articles, book chapters, theses, books, reports) by keyword and metadata. `keyword` is a single substring match over title + abstract, so search ONE term per call (combined terms like 'pèlerinage Mecque' miss results). References are multilingual: try French and English title/abstract keywords when relevant; metadata/filter values such as `reference_type` and `language` use French labels. Results include a short abstract snippet — use get_reference for the full abstract and bibliographic detail.

  Parameters:

  • limit — integer — Default 20, max 100
  • author — string
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional)
  • date_to — string — Latest year, YYYY
  • keyword — string — One French or English concept keyword; substring match on title + abstract (one term per call, accent-insensitive)
  • subject — string — Subject tag (sparse: ~27% of references are tagged)
  • language — string — e.g. Français | Anglais
  • date_from — string — Earliest year, YYYY
  • reference_type — string — Substring match. Values: Article de revue | Chapitre de livre | Livre | Mémoire de maitrise | Rapport | Thèse de doctorat | Communication scientifique | Compte rendu de livre | Article d'encyclopédie | Mémoire de licence | Article de blog | Working paper. Use the full label for precision — 'Livre' alone also matches 'Chapitre de livre' and 'Compte rendu de livre'.

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_reference read-only

  Full bibliographic record for one academic reference (by id), including the complete abstract (present for ~51% of references), subjects, DOI/URL, and host-work details (book, volume, issue, pages).

  Parameters:

  • reference_id (required) — integer

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• semantic_search_references read-only

  Semantic search over academic references by title + abstract, across languages. Only references with an abstract (about half) are covered; use search_references for the rest. Runs the corpus model, EmbeddingGemma 2, locally with no API key (first use downloads ~350 MB). similarity_score only ranks within one query.

  Parameters:

  • query (required) — string — Natural-language query, any language
  • limit — integer — Default 10, max 50
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional)
  • date_to — string — Latest year, YYYY
  • language — string — e.g. Français | Anglais
  • date_from — string — Earliest year, YYYY
  • reference_type — string — French type label, substring (e.g. Thèse de doctorat)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_documents read-only

  Search the small archival-documents subset (~26 items: Islamic association reports, flyers, project documents — mostly Burkina Faso). Use French concept keywords regardless of the user's report language. Most have OCR text and an AI description. Call with no arguments to list all.

  Parameters:

  • limit — integer — Default 15, max 50
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional). Corpus is mostly Burkina Faso/Togo/Benin
  • keyword — string — Concept keyword; substring match on title, OCR, the French and English AI descriptions, and subject (accent-insensitive)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_document read-only

  Get one archival document (by id): full metadata, AI description, and OCR text, in 25k-char parts (follow `next_offset`). Pass a `keyword` to get ~2000-char excerpts around each match instead, useful for long documents.

  Parameters:

  • document_id (required) — integer
  • offset — integer — Read from this character on: the previous part's next_offset
  • keyword — string — Return excerpts around matches instead of the full OCR (accent-insensitive)
  • max_excerpts — integer — Default 10, max 25
  • context_chars — integer — Default 2000, max 5000

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_audiovisual read-only

  Search audiovisual materials by keyword and metadata: francophone web video from Burkina Faso, Togo and Benin (TV reports, association and campus recordings), plus deposited Nigerian Hausa/Arabic recordings. Keyword matches title, creator, publisher, subject, spatial, language, source, the item's own description (the richest text most of these items have) and its transcription where one exists. Each row says which population it is from (`source_type`) and carries either `external_url` (a video to watch) or `media_url` (a file), never both.

  Parameters:

  • limit — integer — Default 20, max 50
  • medium — string — Exact carrier medium: Vidéo sur le web | DVD | CD (validated, accents optional)
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional). Burkina Faso, Togo, Benin and Nigeria only — no Niger or Ivorian items
  • keyword — string — Substring match across audiovisual title/metadata fields
  • subject — string — Exact subject tag — only ~27 rows carry one, so prefer publisher/keyword
  • language — string — Exact language value, e.g. Français | Haoussa | Arabe | Anglais | Mooré
  • publisher — string — Substring on the publishing channel/broadcaster, e.g. RTB | AEEM | CERFI
  • source_type — string — youtube (harvested web video, the large majority) | deposited (recordings with a file, 47)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• list_audiovisual read-only

  List audiovisual materials, newest first (francophone web video from Burkina Faso, Togo and Benin; deposited Nigerian Hausa/Arabic recordings). Filter by country, publishing channel or `source_type`.

  Parameters:

  • limit — integer — Default 20, max 50
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional). Burkina Faso, Togo, Benin and Nigeria only — no Niger or Ivorian items
  • publisher — string — Substring on the publishing channel/broadcaster, e.g. RTB | AEEM | CERFI
  • source_type — string — youtube (harvested web video, the large majority) | deposited (recordings with a file, 47)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_audiovisual read-only

  Get one audiovisual record by id: full description and transcription (where one exists), creator/publishing channel, duration, medium, subjects, places, language, rights, source, and three distinct links: `url` (the IWAC page, the one to cite), `external_url` (where a harvested video plays) and `media_url` (a deposited file). `source_type` says which to expect. A long transcription comes in 25k-char parts (follow `next_offset`); `keyword` returns excerpts instead.

  Parameters:

  • audiovisual_id (required) — integer
  • offset — integer — Read from this character on: the previous part's next_offset
  • keyword — string

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• search_images read-only

  Search the IWAC photographs (30 items: mosques, radio stations, schools, signage and street scenes documented during fieldwork). Keyword matches title, creator, subject, place and the rare caption. Each result carries `image_url` (the full-resolution file), `coordinates` ('lat, lng' where known) and the canonical IWAC page. Call with no arguments to list all. Captions are almost never present, so prefer subject/place filters over keywords, or semantic_search_images when it is enabled.

  Parameters:

  • limit — integer — Default 20, max 50
  • offset — integer
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional)
  • creator — string — Photographer name (substring match)
  • date_to — string — YYYY-MM-DD (or YYYY)
  • keyword — string — French concept keyword; substring match on title, creator, subject, place and caption
  • spatial — string — Exact place name, e.g. Ouagadougou (pipe-aware)
  • subject — string — Exact subject tag (pipe-aware)
  • date_from — string — YYYY-MM-DD (or YYYY)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• get_image read-only

  Get one photograph by id: title, photographer, capture date, place and coordinates, subjects, rights, the IIIF manifest, and the full-resolution `image_url`. The server returns URLs, not image bytes.

  Parameters:

  • image_id (required) — integer

  Hints: read-only yes · destructive no · idempotent yes · open-world no

• semantic_search_images read-only

  Find photographs by describing what they SHOW, in any language ('mosquée en construction', 'street signage in Arabic'). This is cross-modal: the ranking runs against a multimodal embedding of the photograph itself, not against a caption, so it works even though only 2 of the 30 images have one. Runs the corpus model, EmbeddingGemma 2, locally with no API key (first use downloads ~350 MB). similarity_score only ranks within one query.

  Parameters:

  • query (required) — string — Description of the visual content, any language
  • limit — integer — Default 10, max 30
  • country — string — Exact country name: Benin | Burkina Faso | Côte d'Ivoire | Niger | Nigeria | Togo (accents optional)

  Hints: read-only yes · destructive no · idempotent yes · open-world no

Publisher website Repository

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.