MAKE GUIDES ACTUALLY GOOD

CHANGELOG.md

@@ -1,5 +1,14 @@
 # Changelog
 
+## 0.6.3
+
+Generated from the API document of 2026-11-09.
+
+- Transcript responses expose `base_language` and `translation_languages`.
+- Search passages expose `match` (`phrase`, `all_words`, or `partial`) and nullable `line_start_seconds`. Watch links target the matching line when available.
+- Refresh reference descriptions for search filters, transcript windows, alert delivery, and temporary service errors.
+- Update package descriptions and README introductions for transcript search and monitoring.
+
 ## 0.6.2
 
 - Lead the README with tested examples for transcript reads, quotes with source links, and company monitoring.

README.md

@@ -1,6 +1,6 @@
 # Arcmira for Python
 
-Pull YouTube transcripts, search spoken passages, and monitor company mentions with Arcmira. Results include timestamps and source links for research, citations, and agent workflows.
+YouTube transcript search and monitoring for your AI: exact quotes, full transcripts and speakers.
 
 [Arcmira](https://arcmira.com) · [Get an API key](https://arcmira.com/docs/authentication) · [API docs](https://arcmira.com/docs) · [Pricing](https://arcmira.com/pricing)
 
@@ -154,6 +154,8 @@ Entity filters take ids. Resolve names first; explain a `suggested` match and as
 
 ## Use Arcmira with an AI agent
 
+Add "use arcmira" to any question to send it to Arcmira first.
+
 For tasks such as "pull YouTube transcripts", "find quotes with timestamps", "research podcast sponsors", or "monitor company mentions", connect the [Arcmira MCP server](https://github.com/arcmira/mcp). The [agent index](https://arcmira.com/llms.txt) links to current API and citation guidance.
 
 ## Reference

VERSION

@@ -1 +1 @@
-0.6.2
+0.6.3

pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "hatchling.build"
 [project]
 name = "arcmira"
 dynamic = ["version"]
-description = "YouTube and podcast transcript search with the official Arcmira Python SDK. Find timestamped quotes, mentions, sponsors and recommendations."
+description = "YouTube transcript search and monitoring for your AI: exact quotes, full transcripts and speakers."
 readme = "README.md"
 license = "Apache-2.0"
 license-files = ["LICENSE"]

src/arcmira/__init__.py

@@ -246,6 +246,7 @@
         TranscriptResult_Pending,
         TranscriptResult_Ready,
         TranscriptSearchChunk,
+        TranscriptSearchChunkMatch,
         TranscriptSearchResponse,
         TranscriptSearchResponseAccess,
         TranscriptSearchResponseAccessDetails,
@@ -582,6 +583,7 @@
     "TranscriptResult_Pending": ".types",
     "TranscriptResult_Ready": ".types",
     "TranscriptSearchChunk": ".types",
+    "TranscriptSearchChunkMatch": ".types",
     "TranscriptSearchResponse": ".types",
     "TranscriptSearchResponseAccess": ".types",
     "TranscriptSearchResponseAccessDetails": ".types",
@@ -913,6 +915,7 @@ def __dir__():
     "TranscriptResult_Pending",
     "TranscriptResult_Ready",
     "TranscriptSearchChunk",
+    "TranscriptSearchChunkMatch",
     "TranscriptSearchResponse",
     "TranscriptSearchResponseAccess",
     "TranscriptSearchResponseAccessDetails",

src/arcmira/_package.py

@@ -1,5 +1,5 @@
 # Written by scripts/install-generated.py from VERSION.
-__version__ = '0.6.2'
+__version__ = '0.6.3'
 homepage = "https://arcmira.com"
 docs = "https://arcmira.com/docs"
 api_base = "https://api.arcmira.com/v1"

src/arcmira/channels/client.py

@@ -36,7 +36,7 @@ def coverage(
         self, channel_id: str, *, request_options: typing.Optional[RequestOptions] = None
     ) -> ChannelCoverageResponse:
         """
-        How many videos of a YouTube channel are searchable, the newest publish date among them, and the split by transcript source class. Call it when a search or mention lookup came back empty, before telling anyone we do not cover a show, and cite indexed_through as the as-of date for mentions and search_indexed_through for transcript search. It cannot request indexing; channel backfill is not available yet. Free; uses no credits. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
+        How many videos of a YouTube channel are searchable, the newest publish date among them, and the split by transcript source class. Call it when a search or mention lookup came back empty, before telling anyone we do not cover a show, and cite indexed_through as the as-of date for mentions and search_indexed_through for transcript search. It cannot request indexing; to ask for a channel, send feedback with its channel id. Free; uses no credits. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
 
         Parameters
         ----------
@@ -104,7 +104,7 @@ async def coverage(
         self, channel_id: str, *, request_options: typing.Optional[RequestOptions] = None
     ) -> ChannelCoverageResponse:
         """
-        How many videos of a YouTube channel are searchable, the newest publish date among them, and the split by transcript source class. Call it when a search or mention lookup came back empty, before telling anyone we do not cover a show, and cite indexed_through as the as-of date for mentions and search_indexed_through for transcript search. It cannot request indexing; channel backfill is not available yet. Free; uses no credits. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
+        How many videos of a YouTube channel are searchable, the newest publish date among them, and the split by transcript source class. Call it when a search or mention lookup came back empty, before telling anyone we do not cover a show, and cite indexed_through as the as-of date for mentions and search_indexed_through for transcript search. It cannot request indexing; to ask for a channel, send feedback with its channel id. Free; uses no credits. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
 
         Parameters
         ----------

src/arcmira/channels/raw_client.py

@@ -29,7 +29,7 @@ def coverage(
         self, channel_id: str, *, request_options: typing.Optional[RequestOptions] = None
     ) -> HttpResponse[ChannelCoverageResponse]:
         """
-        How many videos of a YouTube channel are searchable, the newest publish date among them, and the split by transcript source class. Call it when a search or mention lookup came back empty, before telling anyone we do not cover a show, and cite indexed_through as the as-of date for mentions and search_indexed_through for transcript search. It cannot request indexing; channel backfill is not available yet. Free; uses no credits. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
+        How many videos of a YouTube channel are searchable, the newest publish date among them, and the split by transcript source class. Call it when a search or mention lookup came back empty, before telling anyone we do not cover a show, and cite indexed_through as the as-of date for mentions and search_indexed_through for transcript search. It cannot request indexing; to ask for a channel, send feedback with its channel id. Free; uses no credits. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
 
         Parameters
         ----------
@@ -143,7 +143,7 @@ async def coverage(
         self, channel_id: str, *, request_options: typing.Optional[RequestOptions] = None
     ) -> AsyncHttpResponse[ChannelCoverageResponse]:
         """
-        How many videos of a YouTube channel are searchable, the newest publish date among them, and the split by transcript source class. Call it when a search or mention lookup came back empty, before telling anyone we do not cover a show, and cite indexed_through as the as-of date for mentions and search_indexed_through for transcript search. It cannot request indexing; channel backfill is not available yet. Free; uses no credits. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
+        How many videos of a YouTube channel are searchable, the newest publish date among them, and the split by transcript source class. Call it when a search or mention lookup came back empty, before telling anyone we do not cover a show, and cite indexed_through as the as-of date for mentions and search_indexed_through for transcript search. It cannot request indexing; to ask for a channel, send feedback with its channel id. Free; uses no credits. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
 
         Parameters
         ----------

src/arcmira/channels/videos/client.py

@@ -36,7 +36,7 @@ def list(
         request_options: typing.Optional[RequestOptions] = None,
     ) -> SyncPager[ChannelVideosResponseEpisodesItem, ChannelVideosResponse]:
         """
-        The indexed videos of a YouTube channel, newest first, each with its video_id, title, publish date, duration, view count, and watch_url on arcmira.com. Pass next_cursor as cursor to continue. The signed token binds the route, filters, caller and visibility; invalid or old tokens return invalid_cursor. A first-page media ID fence excludes later insertions, including old-date backfills; edits and deletions to existing videos remain live. Call it for the latest or most recent episode of a show, or to list what a show published in a window, then pass a video_id to GET /v1/mentions/counts video_ids for what that episode mentions or to GET /v1/transcripts/{video_id} to read it. indexed_through is the newest date we hold for the channel. An empty list means nothing is indexed; channel backfill is not available yet. Uses 4 credits per video returned past the first 5. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
+        The indexed videos of a YouTube channel, newest first, each with its video_id, title, publish date, duration, view count, and watch_url on arcmira.com. Pass next_cursor as cursor to continue. The signed token binds the route, filters, caller and visibility; invalid or old tokens return invalid_cursor. A first-page media ID fence excludes later insertions, including old-date backfills; edits and deletions to existing videos remain live. Call it for the latest or most recent episode of a show, or to list what a show published in a window, then pass a video_id to GET /v1/mentions/counts video_ids for what that episode mentions or to GET /v1/transcripts/{video_id} to read it. indexed_through is the newest date we hold for the channel. An empty list means nothing is indexed; this route cannot add a channel. Uses 4 credits per video returned past the first 5. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
 
         Parameters
         ----------
@@ -110,7 +110,7 @@ async def list(
         request_options: typing.Optional[RequestOptions] = None,
     ) -> AsyncPager[ChannelVideosResponseEpisodesItem, ChannelVideosResponse]:
         """
-        The indexed videos of a YouTube channel, newest first, each with its video_id, title, publish date, duration, view count, and watch_url on arcmira.com. Pass next_cursor as cursor to continue. The signed token binds the route, filters, caller and visibility; invalid or old tokens return invalid_cursor. A first-page media ID fence excludes later insertions, including old-date backfills; edits and deletions to existing videos remain live. Call it for the latest or most recent episode of a show, or to list what a show published in a window, then pass a video_id to GET /v1/mentions/counts video_ids for what that episode mentions or to GET /v1/transcripts/{video_id} to read it. indexed_through is the newest date we hold for the channel. An empty list means nothing is indexed; channel backfill is not available yet. Uses 4 credits per video returned past the first 5. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
+        The indexed videos of a YouTube channel, newest first, each with its video_id, title, publish date, duration, view count, and watch_url on arcmira.com. Pass next_cursor as cursor to continue. The signed token binds the route, filters, caller and visibility; invalid or old tokens return invalid_cursor. A first-page media ID fence excludes later insertions, including old-date backfills; edits and deletions to existing videos remain live. Call it for the latest or most recent episode of a show, or to list what a show published in a window, then pass a video_id to GET /v1/mentions/counts video_ids for what that episode mentions or to GET /v1/transcripts/{video_id} to read it. indexed_through is the newest date we hold for the channel. An empty list means nothing is indexed; this route cannot add a channel. Uses 4 credits per video returned past the first 5. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
 
         Parameters
         ----------

src/arcmira/channels/videos/raw_client.py

@@ -38,7 +38,7 @@ def list(
         request_options: typing.Optional[RequestOptions] = None,
     ) -> SyncPager[ChannelVideosResponseEpisodesItem, ChannelVideosResponse]:
         """
-        The indexed videos of a YouTube channel, newest first, each with its video_id, title, publish date, duration, view count, and watch_url on arcmira.com. Pass next_cursor as cursor to continue. The signed token binds the route, filters, caller and visibility; invalid or old tokens return invalid_cursor. A first-page media ID fence excludes later insertions, including old-date backfills; edits and deletions to existing videos remain live. Call it for the latest or most recent episode of a show, or to list what a show published in a window, then pass a video_id to GET /v1/mentions/counts video_ids for what that episode mentions or to GET /v1/transcripts/{video_id} to read it. indexed_through is the newest date we hold for the channel. An empty list means nothing is indexed; channel backfill is not available yet. Uses 4 credits per video returned past the first 5. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
+        The indexed videos of a YouTube channel, newest first, each with its video_id, title, publish date, duration, view count, and watch_url on arcmira.com. Pass next_cursor as cursor to continue. The signed token binds the route, filters, caller and visibility; invalid or old tokens return invalid_cursor. A first-page media ID fence excludes later insertions, including old-date backfills; edits and deletions to existing videos remain live. Call it for the latest or most recent episode of a show, or to list what a show published in a window, then pass a video_id to GET /v1/mentions/counts video_ids for what that episode mentions or to GET /v1/transcripts/{video_id} to read it. indexed_through is the newest date we hold for the channel. An empty list means nothing is indexed; this route cannot add a channel. Uses 4 credits per video returned past the first 5. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
 
         Parameters
         ----------
@@ -199,7 +199,7 @@ async def list(
         request_options: typing.Optional[RequestOptions] = None,
     ) -> AsyncPager[ChannelVideosResponseEpisodesItem, ChannelVideosResponse]:
         """
-        The indexed videos of a YouTube channel, newest first, each with its video_id, title, publish date, duration, view count, and watch_url on arcmira.com. Pass next_cursor as cursor to continue. The signed token binds the route, filters, caller and visibility; invalid or old tokens return invalid_cursor. A first-page media ID fence excludes later insertions, including old-date backfills; edits and deletions to existing videos remain live. Call it for the latest or most recent episode of a show, or to list what a show published in a window, then pass a video_id to GET /v1/mentions/counts video_ids for what that episode mentions or to GET /v1/transcripts/{video_id} to read it. indexed_through is the newest date we hold for the channel. An empty list means nothing is indexed; channel backfill is not available yet. Uses 4 credits per video returned past the first 5. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
+        The indexed videos of a YouTube channel, newest first, each with its video_id, title, publish date, duration, view count, and watch_url on arcmira.com. Pass next_cursor as cursor to continue. The signed token binds the route, filters, caller and visibility; invalid or old tokens return invalid_cursor. A first-page media ID fence excludes later insertions, including old-date backfills; edits and deletions to existing videos remain live. Call it for the latest or most recent episode of a show, or to list what a show published in a window, then pass a video_id to GET /v1/mentions/counts video_ids for what that episode mentions or to GET /v1/transcripts/{video_id} to read it. indexed_through is the newest date we hold for the channel. An empty list means nothing is indexed; this route cannot add a channel. Uses 4 credits per video returned past the first 5. Every gate is a typed error whose error.unlock.url names the plan that lifts it; pass src=mcp-tool only from the Arcmira MCP server.
 
         Parameters
         ----------

src/arcmira/core/client_wrapper.py

@@ -33,7 +33,7 @@ def get_headers(self) -> typing.Dict[str, str]:
         import platform
 
         headers: typing.Dict[str, str] = {
-            "User-Agent": "arcmira/0.6.2",
+            "User-Agent": "arcmira/0.6.3",
             "X-Fern-Language": "Python",
             "X-Fern-Runtime": f"python/{platform.python_version()}",
             "X-Fern-Platform": f"{platform.system().lower()}/{platform.release()}",