cldf-mcp
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@cldf-mcp Which languages are SOV but use prepositions, and are they spread across families?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
cldf-mcp
An MCP server that gives AI assistants grounded access to cross-linguistic data published in the CLDF format. Version 0.1 ships with:
WALS, the World Atlas of Language Structures: 192 typological features (word order, phonology, morphology…) for 2,660 languages, with sources for every datapoint;
Glottolog: the catalog of the world's ~27,000 languoids, their classification, alternate names, endangerment and description status.
Language models know WALS exists but are unreliable on specifics: which value a language has for feature 85A, or how many languages combine SOV order with prepositions. This server returns the actual data, with the release it came from and the reference behind it.
What makes it different from a lookup table
Family-aware counts. Related languages inherit features, so "400 languages do X" can mean "three big families do X". Every count comes with the number of distinct families (from Glottolog) and WALS genera, plus the share of the largest family, so the model can tell a cross-linguistic tendency from the signature of one lineage. WALS's balanced 100- and 200-language samples are available as filters.
Provenance everywhere. Each answer names the dataset, release and citation; datapoints carry their bibliographic sources and page numbers.
No silent guessing.
bsqis WALS's code for Basque and the ISO code for Bassa, so the server asks instead of picking one. "No data" is never reported as "absent".Pinned, verified data. Datasets are fetched from tagged releases and checked against SHA-256 checksums before loading.
Related MCP server: U.S. Census Bureau Data API MCP
Quick start
Requires Python 3.11+ and uv (or pip).
Claude Code
claude mcp add cldf -- uvx cldf-mcpClaude Desktop and other clients that use an mcpServers config:
{
"mcpServers": {
"cldf": {
"command": "uvx",
"args": ["cldf-mcp"]
}
}
}On first use the server downloads the pinned releases (about 40 MB) and builds a local SQLite database (about 35 MB) in a few seconds. To do this ahead of time:
uvx cldf-mcp setupThen ask things like:
Which languages in WALS are SOV but use prepositions? Are they spread across families?
Is "VSO implies prepositions" a real tendency? Check it within the WALS 200-language sample.
Give me Basque's word-order features, with sources.
How does WALS cover the Austronesian family? Which branches have the most data?
Tools
Tool | What it does |
| Loaded datasets, release versions, coverage, licenses and citations |
| Names (including alternate names and misspellings), ISO 639-3 codes, Glottocodes or dataset codes → ranked candidates with family, macroarea and which entries have data |
| Keyword search over feature names, chapters and value labels → feature IDs and possible values |
| Definition, chapter, authors and value distribution, counted by languages, families and genera; optional filters by macroarea, family/subgroup or WALS sample |
| Language × feature grid with values, comments, sources and full citations; unresolvable names are reported with candidates instead of failing the call |
| Everything about one language: Glottolog identity, lineage, endangerment, description status and all feature values |
| Languages matching up to 8 conditions (AND, with |
| Two features cross-tabulated: languages, families, genera and largest-family share per cell, zero cells kept, optional macroarea breakdown |
| Glottolog lineage, children (with how many dataset entries fall under each), sisters |
| Full bibliographic references for the source IDs cited in datapoints |
Resources:
cldf://guide/typology-caveats: how to use counts responsibly (sampling bias, areal effects, coverage, citation). The server's instructions point the model to it.cldf://datasets: versions, licenses and citations of the loaded data.
All tools are read-only and work offline once the data is prepared.
Configuration
Setting | Default |
|
|
| none: extra TOML to add or override datasets (experimental, see below) |
|
|
Other commands: cldf-mcp info (where data lives, whether it is current) and
cldf-mcp setup --force (rebuild).
Adding a dataset (experimental)
Any CLDF StructureDataset whose languages carry Glottocodes can be loaded alongside WALS. Feature tools, family counting and cross-dataset filters work on it automatically:
# my-datasets.toml
[datasets.mydata]
title = "My typological survey"
role = "features"
version = "v1.0"
base_url = "https://raw.githubusercontent.com/me/mydata/v1.0/cldf/" # or: local_path = "/path/to/cldf"
metadata = "StructureDataset-metadata.json"
license = "CC-BY-4.0"
citation = "Me. 2026. My typological survey."
[datasets.mydata.files] # every file the loader needs, with optional sha256 checksums
"StructureDataset-metadata.json" = ""
"languages.csv" = ""
"parameters.csv" = ""
"codes.csv" = ""
"values.csv" = ""cldf-mcp setup --config my-datasets.tomlData, licenses and citation
The server code is MIT-licensed. The data is not part of this package: it is downloaded from the
original releases, which are licensed under CC-BY-4.0.
If you publish results obtained with this server, cite the datasets (see DATA_LICENSES.md):
Dryer, Matthew S. & Haspelmath, Martin (eds.) 2013. The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. https://wals.info (CLDF release v2020.6)
Hammarström, Harald & Forkel, Robert & Haspelmath, Martin & Bank, Sebastian. 2026. Glottolog 5.3. Leipzig: Max Planck Institute for Evolutionary Anthropology. https://glottolog.org
This is an independent project and is not affiliated with the WALS, Glottolog or CLDF teams.
Limitations
WALS reflects the literature up to about 2013, and most features are coded for a few hundred languages. Percentages are of coded languages, not of the world's languages.
Family counts follow Glottolog, which does not recognise proposed macro-families (e.g. Altaic). Of the 17 WALS languages without a Glottocode, 10 are matched to a Glottolog family through WALS's family or genus name; the other 7 keep WALS's own label (e.g. "Niger-Congo").
Glottolog's full bibliography is not bundled; its references are returned as links.
Roadmap
0.2: Grambank, the first dataset that overlaps with WALS, with side-by-side comparison of related features.
0.3: PHOIBLE sound inventories and Lexibank wordlists (
get_inventory,get_forms).
Development
git clone https://github.com/hsouf/cldf-mcp && cd cldf-mcp
python -m venv .venv && . .venv/bin/activate
pip install -e ".[dev]"
pytest # offline tests on small fixtures (tests/fixtures)
pytest -m network # downloads the real pinned releases and checks known WALS counts
ruff check src tests && ruff format --check src testsSee RELEASING.md for publishing to PyPI and the MCP Registry.
Available Tools
10 toolsclassificationBRead-onlyIdempotent
Glottolog classification of a language or group: lineage, children (with how many dataset entries fall under each), sisters, and the number of languages below.
Inputs: language (required), max_children. Returns: object
crosstabARead-onlyIdempotent
Cross-tabulate two features over the languages coded for both. Each cell gives languages, families and genera attesting the combination and the share of its largest family, so you can tell a cross-linguistic tendency from the signature of one big family. Use this to check implicational universals.
Inputs: feature_a (required), feature_b (required), family, sample, +2 more. Returns: object
filter_languagesARead-onlyIdempotent
Find languages matching feature conditions, e.g. SOV order AND postpositions.
Returns match counts against the number of languages coded for all conditions (the denominator), with family and genus counts and the families contributing most matches, plus the matching languages.
Inputs: conditions (required), limit, family, sample, +1 more. Returns: object
find_languageARead-onlyIdempotent
Find languages by name or code. Returns Glottocodes, family, macroarea and which dataset entries have data.
Use this before other tools when a name might be ambiguous (e.g. 'Arabic' is a family; WALS has Arabic varieties). Candidates are ranked: exact codes, exact names, alternate names, then partial matches.
Inputs: query (required), limit. Returns: object
get_featureARead-onlyIdempotent
Describe one feature: definition, chapter and authors, and how its values are distributed, counted by languages, families and genera.
Inputs: feature (required), family, sample, macroarea, +1 more. Returns: object
get_sourceARead-onlyIdempotent
Full bibliographic references for source IDs cited in datapoints.
Inputs: source_ids (required), dataset. Returns: object
get_valuesARead-onlyIdempotent
Look up specific datapoints: the value of each feature for each language, with sources, comments and full citations. Unresolvable names are reported in 'errors' with candidates instead of failing the call.
When several dataset entries share a Glottocode (e.g. Basque and Basque dialects), the main entry is used and the others are listed; pass a dataset ref like 'wals:bso' to pick a specific one.
Inputs: features (required), languages (required). Returns: object
language_profileARead-onlyIdempotent
Everything known about one language: Glottolog identity, classification, endangerment and description status, plus all its feature values per dataset.
Inputs: language (required), limit, dataset. Returns: object
list_datasetsARead-onlyIdempotent
List the loaded datasets with their release versions, coverage, licenses and how to cite them.
Inputs: no named parameters. Returns: object
search_featuresARead-onlyIdempotent
Search typological features by keyword. Returns feature IDs, names, possible values and how many languages are coded for each.
Inputs: query (required), limit, dataset. Returns: object
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 10 tools
Most tools have clearly distinct roles: find_language (name lookup), search_features (keyword search), get_feature (single feature detail), get_values (datapoint lookup), filter_languages (conditional query), crosstab (two-feature tabulation), classification (hierarchy), get_source (citations). There is some overlap between search_features/get_feature and between get_values/language_profile, but the descriptions clarify that the latter pairs operate at different granularities.
Seven tools follow a verb_noun pattern (list_datasets, find_language, search_features, get_feature, get_values, filter_languages, get_source), but three use bare noun names (language_profile, crosstab, classification). The conventions are mixed but still readable and semantically clear.
Ten tools is well-scoped for a typological-database explorer, with each tool covering a distinct query angle (discovery, lookup, filtering, aggregation, classification, sources). No tool feels redundant or padded.
The surface covers discovery (datasets, languages, features), retrieval (values, profiles, sources), and analysis (filtering, crosstab, classification), which is strong for a read-only typology server. Minor gaps exist, such as no direct export/aggregate-statistics tool, but core workflows are fully supported.
Maintenance
Related MCP Connectors
Grounds LLMs in real Latin: Wiktionary lemmas, Morpheus inflections, citations.
Research-backed lint + generation for agent context files (CLAUDE.md, AGENTS.md) + prompt enhancer.
Connects an AI assistant to your Lingard account (lingard.app), an app for learning languages with books, news and flashcards. Create topic word collections and add complete cards: translation, part of speech, pronunciation, synonyms, example sentences with translations, and images. Fill in missing fields on existing cards, rewrite example translations, and pause or resume collections, books and news in your daily study sessions. Every change can be previewed before it is saved, and saved words appear in the Lingard apps on iOS, Android and the web. Requires a Lingard account; sign-in uses OAuth.
Lexpad is a vocabulary notebook with spaced repetition, in 81 languages. Connect it to ChatGPT or Claude and ask the assistant to save a word from your chat, show the words due today or quiz you. Each answer counts as practice and moves the word through its Leitner boxes. Sign-in is OAuth 2.1 with PKCE on a Lexpad page. 20 tools; the ones that overwrite or delete are marked, so the client asks first. Connecting is free. On Lexpad's Free plan the assistant writes 10 cards a day; on Pro it is unlimited, up to 1,000 a day. Guide and example prompts: https://lexpad.app/features/assistants
Related MCP Servers
- AlicenseAqualityBmaintenanceEnables AI assistants to discover, retrieve, and compare official international development indicators from sources such as the World Bank, FAOSTAT, WHO, UNICEF, and IMF, while preserving source identifiers, units, and citations.102MIT
- FlicenseNot gradedqualityAmaintenanceConnects AI assistants with official Census Bureau data sources, letting them list datasets, resolve place names to FIPS codes, and pull aggregate statistics for states, counties, cities, and other geographies. Queries are served from the Census Data API with a local Postgres mirror for faster, more robust search.-
- AlicenseNot gradedqualityAmaintenanceEnables AI assistants to query official Census Bureau statistics by listing datasets, retrieving geographies and aggregate data, and resolving geographic names to FIPS codes through natural-language interactions.MIT
- AlicenseAqualityCmaintenanceEnables LLM-powered applications to look up bacterial and archaeal taxonomy, search genomes, browse and navigate the taxonomic tree, and compare genomes or find phylogenetic paths between taxa. It works out of the box with a built-in reference dataset covering major phyla and can be pointed at full Genome Taxonomy Database TSV files for comprehensive queries.121MIT