msza.ai — Polish Catholic Sermons ✝

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

Search public Polish Catholic sermons, transcripts, Scripture, prayers, churches and daily content.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "msza.ai — Polish Catholic Sermons ✝": {
      "url": "https://msza.ai/mcp",
      "transport": "streamable-http"
    }
  }
}

Measured tools (17)

• search_msza_ai read-only

  Searches public, predominantly Polish-language msza.ai sermons, sermon studies, Bible passages, prayers, songs, people, and places by topic, keyword, and relevance. Results retain stored source language and contain candidate matches, stable source IDs, citation labels, short excerpts, dates, and links. Date and church filters apply to sermon records. Priest names are unavailable.

  Parameters:

  • query (required) — string — Focused Polish retrieval query for a topical request. For another user language, translate only topical terms into Polish while preserving proper names, exact IDs, Bible references, and supplied filters.
  • limit — integer — Maximum number of compact source results to return.
  • date_to — string — Optional last Warsaw calendar day of an absolute archive period, inclusive.
  • weekday — integer — Optional Warsaw weekday filter: Sunday=0, Monday=1, ..., Saturday=6.
  • camera_id — string — Optional stable church profile ID. When supplied, only public sermons from that church are searched.
  • date_from — string — Optional first Warsaw calendar day of an absolute archive period, inclusive.
  • days_back — integer — Optional rolling calendar-day window ending today, for example 7 or 28. Omit when using date_from or date_to.
  • content_types — array — Optional content categories to search; omitted preserves automatic mixed discovery. Unselected categories are not queried. prayer_book covers prayers, songs and fixed liturgical texts. ai_study results are AI-generated orientation, not evidence. Explicit bible selection enables bounded Scripture search without intent inference. Date/weekday/church filters apply only to sermons; camera_id requires sermon when content_types is supplied. Up to 6 entries; duplicates are removed. Existing source kind labels remain unchanged.

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

• get_sermon read-only

  Returns public sermon text for an exact session ID in its stored language. text_type identifies an AI-edited transcript section, fallback summary, or fallback short thought; these forms are not guaranteed verbatim speech. truncated and total_chars describe text completeness. references contains at most 50 saved annotations; selected_section concerns the displayed section and session can concern other liturgy. Saved annotations can be machine-generated and incomplete.

  Parameters:

  • session_id (required) — string — Exact public sermon session ID returned by search_msza_ai or a sermon-listing tool.

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

• get_daily_page read-only

  Returns an already-prepared public /dzisiaj briefing for today in the Europe/Warsaw timezone or one YYYY-MM-DD date. The briefing contains a daily thought, readings, patrons, name days, external homilies, and public sermon links. sermon_aggregates_available identifies whether stored synthesis and digest provenance remains verifiable; false makes zero digest counters placeholders. AI-generated thoughts and synthesis are orientation material. Data comes from stored SQLite content without generation or refresh.

  Parameters:

  • date — string — Optional Warsaw calendar date. Omit it to use today.

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

• list_recent_sermons read-only

  Returns up to 10 latest public msza.ai sermons in reverse chronological order, optionally limited to one church. This bounded, non-paginated listing contains compact metadata and excerpts.

  Parameters:

  • camera_id — string — Optional stable public church profile ID. Omit it to list recent sermons from all churches.

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

• list_sermons_by_date read-only

  Returns public sermons whose start time falls on one selected Europe/Warsaw calendar day, optionally limited to one church. The non-paginated listing contains at most 50 compact records; total and truncated describe the safety ceiling.

  Parameters:

  • date (required) — string — Required Europe/Warsaw calendar date using YYYY-MM-DD.
  • camera_id — string — Optional stable public church profile ID. Omit it to list sermons from all churches on the selected day.

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

• get_prayer read-only

  Returns one saved public prayer-book item: prayer, song, or fixed liturgical text. Bare prayer IDs and namespaced song:<id> and standard:<id> IDs are supported. kind, text_status, and truncated describe content availability; a song can contain only a first line or metadata. rights_note records provenance rather than permission. Text is stored content without generation or network retrieval.

  Parameters:

  • prayer_id (required) — string — Existing bare prayer ID, or exact namespaced prayer_id/item_id (song:<id> or standard:<id>) returned by search or transcript/sermon references.

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

• get_person read-only

  Returns one curated Catholic or biblical person profile by exact public person ID. Session-derived dynamic profiles remain unavailable until public provenance is verifiable. Results contain bounded profile fields, canonical source URLs, and at most 20 indexed Scripture references.

  Parameters:

  • person_id (required) — string — Exact person sourceId returned by search_msza_ai, or id of a person in get_sermon.references.

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

• get_place read-only

  Returns one public Catholic or biblical place profile by exact public place ID. Results contain bounded profile fields, canonical source URLs, and at most 20 indexed Scripture references.

  Parameters:

  • place_id (required) — string — Exact place sourceId returned by search_msza_ai, or id of a place in get_sermon.references.

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

• get_bible_passage read-only

  Returns up to 50 unchanged UBG verses from one chapter with copyright, license, and source attribution. The local text uses USFM book IDs such as GEN, PSA, MAT, and JHN. No translation or AI generation occurs.

  Parameters:

  • book_id (required) — string — USFM book ID, for example JHN for John or GEN for Genesis (case-insensitive).
  • chapter (required) — integer — Chapter number within the selected UBG book.
  • start_verse (required) — integer — First verse, inclusive.
  • end_verse — integer — Last verse, inclusive. Omit to read only start_verse. At most 50 verses per request.

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

• list_churches read-only

  Returns enabled public churches in stable camera ID order, up to 50 per page. next_cursor represents the next page for the same optional literal name/address substring filter; null marks the end. truncated identifies additional matching records. The live catalogue can change during traversal.

  Parameters:

  • limit — integer — Maximum number of churches to return (default 20, maximum 50).
  • query — string — Optional literal substring of a church name or address. Preserve Polish diacritics; case folding is limited to ASCII letters. Keep the same query on subsequent pages.
  • cursor — string — Opaque next_cursor from the previous page. Omit for the first page. Bound to the same query.

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

• get_church read-only

  Returns a saved public church profile by camera_id with bounded description, address, stored Mass schedule, website, and update time. Profiles can be AI-enriched and schedules can be outdated. The result reads saved data without profile refresh or generation.

  Parameters:

  • camera_id (required) — string — Exact public camera_id returned by list_churches.

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

• get_transcript read-only

  Returns saved final transcript segments from an ended or interrupted public session rather than an AI-edited sermon section. Each page contains at most 20 segments and 12000 UTF-16 text units. next_cursor represents the next page for the same session; long segments continue with char_offset. Text is privacy-redacted, timestamps retain original metadata, and creation-time/ID ordering is stable. Pagination is live, so edits can invalidate a cursor. Saved references concern the returned page. Each page consumes the shared listing budget; oversized stored segments fail explicitly.

  Parameters:

  • session_id (required) — string — Exact public session ID from search_msza_ai or a sermon listing. Only ended/interrupted public sessions are accessible.
  • cursor — string — Opaque next_cursor from the previous page of this same session. Omit for the first page.

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

• get_announcements read-only

  Returns saved parish announcements for an ended or interrupted public session, preferring the edited announcements section and otherwise its summary tab. The response excludes sermon and full-recording fallbacks. text_type identifies AI-edited or summarized text rather than guaranteed verbatim speech. Recording and summary-update dates are not announcement expiry dates. Text is limited to 8000 UTF-16 units; truncated and total_chars describe completeness. No AI generation or network retrieval occurs. Shared reference rate limits apply.

  Parameters:

  • session_id (required) — string — Exact public recording session ID, for example from a sermon listing or an msza.ai recording URL. The recording must contain saved announcements.

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

• get_ai_study read-only

  Returns one saved public AI study by study_id with its complete public text and source-sermon metadata. An exceptional safety limit fails explicitly rather than returning partial text. source_sermons contains currently public finished sermons from at most 200 linked candidates. sources_truncated describes the candidate cap rather than text truncation. The response identifies AI-generated origin and includes provenance fields.

  Parameters:

  • study_id (required) — string — Exact sourceId/study_id of an AI study returned by search_msza_ai, or the ID in its canonical msza.ai study URL.

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

• find_sessions_by_bible_passage read-only

  Returns public saved recordings whose saved Bible annotations exactly overlap the requested single-chapter verse range. scope=sermon excludes reading-only annotations; scope=session includes other liturgy annotations. Results use live lexicographic session_id order rather than chronology. Matches are stored annotations rather than semantic search or verified evidence. The indexed scan covers at most 200 public chapter candidates per request. No content generation, Bible-text retrieval, or network access occurs. Each page consumes the shared listing quota.

  Parameters:

  • book_id (required) — string — Three-character USFM book ID (case-insensitive), such as JHN or GEN; Catholic TOB, JDT, WIS, SIR, BAR, 1MA and 2MA are also accepted.
  • chapter (required) — integer — Chapter number.
  • start_verse (required) — integer — First requested verse, inclusive.
  • limit — integer — Maximum recordings to return (default 10, maximum 20).
  • scope — string — sermon excludes reading-only annotations; session includes annotations from other liturgy sections.
  • cursor — string — Opaque next_cursor for the identical passage and scope. Page positions use session_id order, not recording date; additions behind a live cursor require restarting.
  • end_verse — integer — Last requested verse, inclusive; omit for one verse. Requests span at most 50 verses.

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

• list_agent_skills read-only

  Returns up to 50 immutable public Agent Skill summaries from the msza.ai catalogue, optionally filtered by one exact tag. Results contain metadata and canonical links without full Markdown. The catalogue is read-only; publishing belongs to a separate authenticated Hermes endpoint.

  Parameters:

  • tag — string — Optional exact public catalogue tag; omit it to list all published skill heads.
  • limit — integer — Optional page size from 1 to 50; the default is 20.
  • offset — integer — Optional zero-based offset from 0 to 1000 for bounded catalogue pagination.

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

• get_agent_skill read-only

  Returns one immutable public Agent Skill by exact lowercase slug and optional x.y.z version. An omitted version returns the current published head; a supplied version returns that immutable snapshot. The read-only result contains complete skillMarkdown and metadata. Publishing belongs to a separate authenticated Hermes endpoint.

  Parameters:

  • slug (required) — string — Required exact lowercase ASCII kebab-case skill slug returned by list_agent_skills.
  • version — string — Optional exact immutable semantic version x.y.z; omit it to read the current published head.

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

Publisher website

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