Skip to main content
Glama
PrattyT85

theosis-ancient-context-mcp

by PrattyT85

theosis-ancient-context-mcp

FastMCP stdio service for ancient text corpora. Part of the Theosis ancient-context integration.

Corpora

Corpus

Language(s)

Status

Licence

TLA (Thesaurus Linguae Aegyptiae)

Egyptian

Remote/status-only

Free non-commercial research

Coptic SCRIPTORIUM

Coptic

Local optional

CC-BY (exceptions exist)

HPM/HDivT (Hittite ritual texts)

Hittite

Remote/status-only

Academic

CUC (Copenhagen Ugaritic Corpus)

Ugaritic

Local optional

CC BY-NC 4.0

DASI

Arabic epigraphic

Remote/status-only

Academic

OCIANA

Akkadian/Sumerian

Remote/status-only

Academic

CDLI

Akkadian/Sumerian/Elamite

Remote/status-only

Open access

DPPC

Phoenician/Punic

Deferred

Not published

CIP

Punic

Deferred

Not published

Related MCP server: DEFAIR MCP Server

What's integrated vs. what's not

  • DPPC and CIP are NOT integrated — no public API, data, or licence has been confirmed. They are registered as deferred with no scraper or adapter.

  • CUC is CC BY-NC and local-only — commercial use is prohibited. The adapter only works when CUC_CORPUS_DIR is configured to point at a local checkout of the corpus.

  • TLA, HPM, DASI, OCIANA are remote/status-only until a stable public API contract is verified. They return structured remote_only results with provenance envelopes.

  • CDLI is status-only until the REST JSON API shape is confirmed from cdli.earth/docs/api.

  • Coptic SCRIPTORIUM requires a local corpus directory. The adapter supports meta.json-backed metadata and bounded text-file search.

Tools

  1. list_corpora — all registered corpus records with metadata

  2. get_corpus_status(corpus) — detailed status, local path, availability

  3. search_corpus(corpus, query, limit) — search with explicit status results

  4. get_text(corpus, reference) — text retrieval with provenance envelope

  5. get_text_metadata(corpus, reference) — metadata retrieval with provenance

Setup

# Install
uv sync --extra dev

# Run tests
uv run pytest -v

# Compile check
uv run python -m compileall src/theosis_ancient_context -q

# Run the server (stdio)
uv run python -m theosis_ancient_context

Optional local corpora

# Coptic SCRIPTORIUM
export COPTSCRIPTORIUM_CORPUS_DIR=/path/to/corpora

# Copenhagen Ugaritic Corpus
export CUC_CORPUS_DIR=/path/to/cuc

Integration

All results include a provenance envelope with:

  • source_layer — integration type (local_optional, remote_only, deferred, etc.)

  • corpus_id — canonical corpus identifier

  • source_url — upstream URL

  • licence — licence summary

  • licence_warning — detailed licence notes and restrictions

  • access_status — current integration status

Safety

  • Path traversal prevention for all local file access

  • Bounded search results and text retrieval

  • No persistent cache

  • No secrets required

  • No network requests in v0.1.0 (all adapters are local or status-only)

Related MCP Connectors

Related MCP Servers

  • A
    license
    Not graded
    quality
    A
    maintenance
    Enables AI agents to query a local, offline Crimson Desert knowledge base through stdio MCP, providing versioned entity/claim lookups, bounded search, codebook decoding, and typed responses with provenance context.
    Apache 2.0
  • A
    license
    A
    quality
    B
    maintenance
    Exposes forensic case and evidence management operations over stdio, including case creation, evidence registration with SHA-256 hashing, and integrity verification, so clients can manage DFIR investigations programmatically.
    7
    MIT
  • A
    license
    Not graded
    quality
    B
    maintenance
    Enables any agent to query the same interview corpus as the UI over stdio, exposing tools to list transcripts, search quoted lines, retrieve verified themes with citations and dropped reasons, and fetch source lines.
    17 npm
    MIT
  • F
    license
    Not graded
    quality
    C
    maintenance
    Gives an LLM client BM25 keyword search over a local corpus of Markdown and text files, letting it find ranked passages, fetch exact chunks with source offsets, read whole documents, and list what was indexed. It runs entirely locally over stdio with an in-memory index, calling no model and requiring no API key.
    -