African Speech Corpora MCP
This server lets you explore, audit, and plan training data from African speech corpora (Wolof-first, plus Pulaar and Sereer) through read-only MCP tools.
List corpora – View all language/task/config variants, audit status, verification state, recommended use, and whether they are queryable.
Audit a corpus – Get dated local quality metrics (raw/usable/transcribed/audit-accepted/expert-verified counts and durations) separately from published figures, with provenance and warnings.
Plan training sets – Build de-duplicated, task-aware plans for ASR/TTS/KWS/intent using quality levels (any, transcribed, audit_accepted, expert_verified, wolof_only), with license and benchmark filtering.
Search segments – Search rows in schema-verified variants, with provenance and ephemeral audio URLs.
Filter segments – Filter rows when a row-level manifest exists; otherwise return a structured refusal instead of fake exclusions.
Compare corpora – Compare quality, domain, language purity, license, commercial use, and benchmark status across up to 20 variants.
Get corpus stats – Return exact config-scoped row counts, declared splits, observed configs, and schema discrepancies.
Cite corpora – Retrieve license, BibTeX citation, and parent citations for derivatives.
Provides read-only access to African speech corpora hosted on Hugging Face, enabling tools to list corpora, search transcription segments, retrieve corpus statistics, cite sources, and plan de-duplicated training sets with full provenance.
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., "@African Speech Corpora MCP Plan a de-duplicated training set for Wolof"
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.
African Speech Corpora MCP
A read-only server implementing the Model Context Protocol, an open standard created by Anthropic, over public speech corpora for African languages: Wolof first, plus Pulaar and Sereer.
The project is Wolof-first. The 2.0 catalog contains 14 variants: 11 variants of original sources and 3 derivatives. Pulaar (ful) and Sereer (srr) are represented only by Kallaama variants without local metrics; the project does not claim Swahili or Amharic coverage. The bundled validation lock, generated on August 26, 2026, makes 6 Hugging Face variants queryable. A later validation run may naturally produce a different state.
Version 2.0.0 is distributed on PyPI and published as io.github.papasega/african-speech-corpora in the official MCP Registry.
The server does not train models, download complete corpora, or write to Hugging Face, OpenSLR, Kaggle, GitHub, or any other remote source.
What the server measures
The quality pipeline keeps the following stages separate:
raw audio → usable audio → transcribed audio → audit-accepted audio → expert-verified audioRaw: a file present in the observed snapshot.
Usable: a file remaining after the audit's quantifiable exclusions.
Transcribed: audio associated with a transcription.
Audit accepted: an explicitly named union of expert-verified material and material only assumed valid by the audit.
Expert verified: only
expert_auditedorsource_reported_expert. An “a priori” assessment never enters this level.
Seconds are the canonical duration representation. Decimal hours and HH:MM:SS strings are derived at serialization time. Every audited metric states its scope, source, method, observation date, and confidence. A missing value remains null; it is never converted to zero. Figures published by a project remain under published_metrics, separate from local observations.
Related MCP server: Asset Library MCP Server
2026 Wolof snapshot
The machine-readable source is assets/wolof-audit-2026.csv. The protocol, limitations, and discrepancies with the source table's TOTAL cells are documented in assets/wolof-audit-2026.md.
Seven Wolof variants have a local observation: ALFFA, FLEURS wo_sn, Kallaama Wolof, Urban Bus, Waxal crowdsource, Wolof TTS Baamtu, and WolBanking77. Totals are recomputed from these seven rows and are never stored as a redundant manual total: 148,102 usable files, 64,609 transcribed files, 50,494 audit-accepted files, 2,329,382.06 seconds of audio, 515,185.06 transcribed seconds, and 302,048.06 audit-accepted seconds.
Important limitations:
the FLEURS, Urban Bus, Waxal, Wolof TTS, and WolBanking77 observations described as “a priori” are
audit_assumed_valid, not expert verified;ALFFA explicitly has zero expert-verified files in this snapshot;
the 153 Kallaama files are long radio or interview recordings, not 153 speech turns; 36 files and 12:49:36 are attributed to the local protocol's
source_reported_expertbasis;Urban Bus contains substantial French content without a quantified rate and therefore has
language_purity=mixed_fr;Waxal separates 517:38:05 of raw audio, usable notably for SSL, from only 13:41:28 transcribed for supervised ASR;
the local Wolof TTS Baamtu snapshot—36,009 files and 37:04:49—is distinct from the rounded public metric, and this TTS corpus is excluded from ASR by default;
WolBanking77 distinguishes 2,563 observed audio files from the 9,791 text phrases reported elsewhere;
Afrivoice publishes 530.74 hours of Wolof audio, including 102.96 transcribed hours. These are source-published figures, not measurements from the 2026 local audit and not evidence of expert verification. The dataset is auto-gated on Hugging Face, so its files, schema, splits, and durations cannot be independently checked without accepting its access conditions and supplying a token; it therefore remains
audit_status=pending;because the exact source observation date is unavailable,
observed_at=2026intentionally has year-only precision.
Installation
The package requires Python 3.11 or newer. Python 3.12 is recommended and is selected explicitly below so that the virtual environment does not accidentally inherit an older system interpreter such as Python 3.9:
Install the published release from PyPI:
python3.12 --version
python3.12 -m venv .venv
source .venv/bin/activate
python --version
python -m pip install "african-speech-mcp==2.0.0"For development, install an editable checkout with the development dependencies:
git clone https://github.com/papasega/african-speech-mcp.git
cd african-speech-mcp
python3.12 --version
python3.12 -m venv .venv
source .venv/bin/activate
python --version
python -m pip install -e ".[dev]"Both version commands should report Python 3.12.x. Creating the environment with python -m venv .venv is safe only when that python executable is already Python 3.11 or newer. If installation fails with an error such as:
ERROR: Package 'african-speech-mcp' requires a different Python: 3.9.6 not in '>=3.11'then the virtual environment was created with Python 3.9.6. Deactivate it, remove or rename that local .venv, install Python 3.12 if necessary, and recreate the environment with python3.12 -m venv .venv. A virtual environment keeps the interpreter with which it was created; activating it does not upgrade Python.
The stdio server starts with no arguments:
african-speech-mcpIt can also be started explicitly:
african-speech-mcp serve --transport stdio
african-speech-mcp serve --transport streamable-httpExample Claude Desktop configuration after installing the package, preferably using an absolute path:
{
"mcpServers": {
"african-speech-corpora": {
"command": "/absolute/path/to/.venv/bin/african-speech-mcp"
}
}
}Version 2.0.0 is published on PyPI and can be installed with pip. The repository installation remains the appropriate choice for development or unreleased changes.
MCP tools
All eight tools declare read_only_hint=true, destructive_hint=false, and idempotent_hint=true.
Tool | Purpose |
| List variants, audit state, recommended use, and validation state. |
| Return counts, durations, verification basis, warnings, and separate published figures. |
| Build a plan without double-counting derivatives or including benchmarks by default. |
| Filter only when a row-level manifest exists; otherwise refuse to promise exclusions. |
| Compare quality, domain, language purity, license, and metric freshness. |
| Search rows in a variant whose remote schema has been validated. |
| Return sizes for the variant's config only and report schema discrepancies. |
| Return the license, citation, and parent citations for a derivative. |
Conceptual examples:
audit_corpus(corpus="waxal-crowdsource")
plan_training_set(language="wol", task="asr", quality="transcribed")
plan_training_set(language="wol", task="asr", quality="expert_verified")
plan_training_set(
language="wol",
task="asr",
quality="audit_accepted",
include_mixed_language=true,
)
filter_segments(
corpus="waxal-crowdsource",
exclude_duplicates=true,
exclude_non_wolof=true,
exclude_corrupt=true,
)The last call currently returns filter_available=false. Aggregate totals establish that Waxal contains 430 duplicates and 22 corrupt or non-Wolof files, but they provide no row identifiers. The server therefore refuses to pretend that it removed those rows.
plan_training_set semantics
quality accepts:
any: select by task and license without requiring transcription;transcribed: use actually transcribed duration;audit_accepted: include expert and assumed-valid material, with a visible breakdown and warning;expert_verified: use only genuinely expert verification bases;wolof_only: excludemixed_fr,mixed,unknown, and variants without sufficient language-purity evidence.
metric_source is either latest_audit (the default) or published. Variants without a known duration remain listed in hours_unknown_for. FLEURS is a benchmark and stays out of training unless include_benchmarks=true. Urban Bus requires include_mixed_language=true. TTS variants stay out of ASR unless explicitly enabled, and derivatives are never added to their parents. Unknown, non-commercial, share-alike, or unconfirmed licenses produce appropriate warnings; commercial_use=null is never presented as commercially compatible.
Persistent Hugging Face validation
african-speech-mcp validate
african-speech-mcp validate --write
african-speech-mcp validate --output validation-lock.jsonValidation checks the dataset identifier, config, splits, transcription column, and any declared language column. The lock is written through atomic replacement and records checked_at, observed schema information, and a stable fingerprint. It is also bound to the catalog hash, so the server rejects a stale lock.
By default, --write creates validation-lock.json in the current directory without modifying the installed package. To use it afterward:
export ASM_VALIDATION_LOCK_PATH="$PWD/validation-lock.json"
african-speech-mcpPossible states are unverified, verified, gated, unavailable, schema_mismatch, and quality_blocked. A responding identifier is never marked verified when the expected config or column is missing. For Afrivoice, set ASM_HF_TOKEN locally before validation; the token is neither serialized nor logged.
corpus_stats filters strictly by config, so FLEURS wo_sn never includes another language. A multilingual distribution without a separate config must declare a language column and accepted values. When the server cannot guarantee that filter, it refuses the request or returns an explicit warning.
Catalog and architecture
src/african_speech_mcp/
├── audit.py # audited CSV, duration formats, deterministic aggregation
├── catalog.py # Pydantic v2 variants and invariants
├── cli.py # serve, validate, catalog-show
├── config.py # ASM_* settings
├── hf_client.py # GET only, retries, bounded concurrency
├── models.py # structured responses
├── server.py # eight MCP tools
├── validation.py # schema inspection and atomic lock files
└── data/
├── corpora.json # single canonical catalog
└── validation-lock.jsonThere is no longer a second root-level corpora.json. The audit CSV remains canonical under assets/, and Hatch includes it deterministically in the wheel. Tests check totals and references between the CSV and catalog.
Useful variables include ASM_CATALOG_PATH, ASM_VALIDATION_LOCK_PATH, ASM_AUDIT_LOG_PATH, ASM_HF_TOKEN, ASM_REQUEST_TIMEOUT_S, ASM_REQUEST_RETRIES, and ASM_VALIDATION_CONCURRENCY. Runtime logs go to stderr so that stdout remains reserved for the stdio protocol.
Contributing an audit or manifest
A new audit must:
add unit rows to the CSV, never a manual TOTAL row;
provide
audit_id,observed_at,source_reference, method, confidence, and verification basis;distinguish an observed zero from a missing value;
reference the audit from the matching variant;
document limitations in Markdown and add consistency tests.
To enable real filtering, provide a versioned manifest with at least:
audit_id,corpus_key,revision,row_id,audio_path,is_duplicate,is_empty,is_corrupt,is_wolof,transcript_quality,reasonIdentifiers must come from the audited snapshot. They must never be invented from aggregate counts.
Development and release preparation
python -m pip install -e ".[dev]"
ruff check src tests scripts
pytest -q
python -m buildCI runs these checks on Python 3.11 and 3.12, installs the built wheel in a clean environment, and tests both the CLI and a real MCP stdio exchange. Version 2.0.0 is a documented breaking release: the public unit is now a single-language, single-task, single-config variant with more structured quality responses.
Publishing a non-prerelease GitHub Release runs .github/workflows/publish-pypi.yml. The workflow verifies the attached distributions, publishes them to PyPI through Trusted Publishing, then validates and publishes the matching server.json to the official MCP Registry through GitHub OIDC. No long-lived PyPI or MCP Registry token is stored. If PyPI succeeded but the registry step failed, rerun the workflow manually with the same tag and registry_only=true; an already active registry version is detected and treated as success.
License
The server code is licensed under Apache-2.0. Each corpus retains its own license. Catalog entries marked TO BE CONFIRMED or SEE DATASET ... intentionally remain uncertain and must be checked against the primary source before any use, especially commercial use.
Available Tools
8 toolsaudit_corpusARead-onlyIdempotent
Return dated observed quality metrics separately from published figures. Missing observations remain null, never implicit zero.
Inputs: corpus (required), audit_id, include_published. Returns: task, config, corpus, counts, +22 more
cite_corpusBRead-onlyIdempotent
Return licence and citations, including parents of derivatives.
Inputs: corpus (required). Returns: bibtex, licence, homepage, also_cite, +2 more
compare_corporaBRead-onlyIdempotent
Compare variant quality, provenance, licence and benchmark status.
Inputs: corpora (required). Returns: corpora, compared, warnings
corpus_statsARead-onlyIdempotent
Return rows and schema for exactly one declared dataset configuration.
Inputs: corpus (required). Returns: splits, columns, warnings, provenance, +4 more
filter_segmentsCRead-onlyIdempotent
Filter rows only when a versioned row-level audit manifest exists; aggregated audit totals alone produce a structured refusal.
Inputs: corpus (required), limit, query, split, +4 more. Returns: corpus, reason, audit_id, segments, +5 more
list_corporaBRead-onlyIdempotent
List language/task/config variants with audit and verification state.
Inputs: task, language, include_derivatives. Returns: total, corpora, disclaimer, audited_count, +3 more
plan_training_setARead-onlyIdempotent
Build a de-duplicated variant plan using published or dated audit metrics, with explicit quality, licence, benchmark and language-purity decisions.
Inputs: task, quality, language, metric_source, +4 more. Returns: task, quality, included, language, +14 more
search_segmentsCRead-onlyIdempotent
Search a schema-verified variant. Config overrides cannot cross language boundaries; temporary audio URLs are explicitly marked ephemeral.
Inputs: query (required), corpus (required), limit, split, +1 more. Returns: query, returned, segments, warnings, +3 more
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 8 tools
Each tool targets a distinct operation: listing variants, searching segments, getting stats, citing, auditing, planning, filtering, and comparing. Though search_segments and filter_segments both work on segments, their preconditions and purposes are clearly different—search is free-form with overrides, while filter requires a versioned audit manifest. No two tools appear interchangeable.
Most tools follow a clear verb_noun pattern (list_corpora, search_segments, cite_corpus, audit_corpus, plan_training_set, filter_segments, compare_corpora). The outlier is 'corpus_stats' which uses noun_noun instead of a get/stats verb, breaking the pattern slightly. This is a minor deviation that doesn't hinder readability.
With 8 tools, the server is well-scoped for its domain of managing and analyzing African speech corpora. Each tool addresses a distinct aspect of the workflow without redundancy or bloat. The count feels just right—not too thin, not overwhelming.
The tool surface covers the core lifecycle of corpus exploration and usage: discovering corpora, searching segments, retrieving stats, citations, audit metrics, planning training sets, filtering based on audits, and comparing variants. There are no glaring gaps such as missing download or update operations, as the focus is analysis rather than modification. The server fulfills its stated purpose comprehensively.
Maintenance
Related MCP Connectors
African-language translation, chat, speech-to-text and text-to-speech from Sunbird AI.
Search speech in podcasts, government meetings, and your own audio: speakers, entities, timestamps.
Read-only access to the Islam West Africa Collection via Hugging Face datasets.
Read-only corpus catalogs and bounded observations with source dates and rights metadata.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceAn MIT-licensed MCP server that provides access to a rolling 24-month corpus of NSE/BSE company documents (concall transcripts, annual reports, investor presentations) with provenance, deduplication, and rights management. It enables searching companies, finding documents, retrieving document content, and checking coverage through a stateless Streamable HTTP MCP endpoint.MIT
- FlicenseNot gradedqualityDmaintenanceEnables users to search and retrieve research datasets via natural language queries, with tools for proposing and committing tag changes. Includes robust authorization, idempotent mutations, and an audit trail.-

corpus.333.ecoofficial
FlicenseNot gradedqualityBmaintenanceEnables agents to search, retrieve, and list open-licensed documents with verifiable provenance, attaching sha256, DOI, and OpenTimestamps proof to every response.1,531 npm-- AlicenseNot gradedqualityBmaintenanceProvides read-only search and context-pack creation over a local source library, letting AI assistants retrieve relevant excerpts and audit cited quotations.20MIT