> ## Documentation Index > Fetch the complete documentation index at: https://docs.arrays.org/llms.txt > Use this file to discover all available pages before exploring further. # Get Podcast Transcripts > [OpenAPI JSON Spec](/docs/output/v1_other_podcast_transcripts_get.json) Retrieve stored official podcast transcripts, newest first. Filter by `podcast_show_id`, `podcast_show_name` or `speaker` — at least one is required unless both `updated_after` and `updated_before` are supplied. Update polling selects episode rows within [updated_after, updated_before), ordered by updated_at ASC, id ASC. Times are Unix seconds; all shows may be queried with a bounded window. Existing filters intersect that window. Updates include episode metadata, transcript and attribution changes, not lease/retry bookkeeping. Show/person metadata has a separate lifecycle. Each item returns the stable Arrays record `id`, the publisher `episode_id`, and RFC3339 `updated_at`. For update polling, pass pagination.cursor unchanged with the same window, and filters. Limit and include_raw may change on retry without restarting. An omitted cursor with has_more=false means the window is drained. Start each new window without the old cursor. Offset is not allowed. This is current-state polling, not a snapshot or deletion/history feed: overlap windows and reconcile delayed writes; intermediate versions may coalesce. `podcast_show_id` comes from `/other/podcast/shows`. `podcast_show_name` takes the name instead and resolves it server-side. The response reports the show name as `podcast_show_name`. `speaker` matches a person in the returned `resolved_speakers`, including host, guest and clip attributions. Registry names and aliases match linked identities; stored resolved names also match directly. Publisher-only labels and attributions excluded from the response do not match. Clips are not guests. `resolved_speakers[*].person_id`, `hosts` and `guests` contain only terminal person-registry UUIDs, matching `/persons`. Historical IDs and merged UUIDs are resolved on read. Unlinked speakers keep their name but omit `person_id`; unresolved references are omitted from `hosts` and `guests`. ## OpenAPI ````yaml /api-reference/openapi.json get /v1/other/podcast/transcripts openapi: 3.0.0 info: contact: {} description: >- Welcome to Arrays! Arrays is a unified financial data layer for both financial institutions and retail users, delivering comprehensive market data across crypto and equity markets. Our datasets cover fundamentals, ETFs, options, crypto, and more. Visit [arrays.org](https://arrays.org) to request an API key and start your free trial. You can explore our full API offerings in the documentation below. For any questions regarding APIs or pricing, please contact support@arrays.org. title: Arrays API version: '1.0' servers: - url: https://data-tools.prd.arrays.org/api security: [] paths: /v1/other/podcast/transcripts: get: tags: - Other summary: Get Podcast Transcripts description: >- [OpenAPI JSON Spec](/docs/output/v1_other_podcast_transcripts_get.json) Retrieve stored official podcast transcripts, newest first. Filter by `podcast_show_id`, `podcast_show_name` or `speaker` — at least one is required unless both `updated_after` and `updated_before` are supplied. Update polling selects episode rows within [updated_after, updated_before), ordered by updated_at ASC, id ASC. Times are Unix seconds; all shows may be queried with a bounded window. Existing filters intersect that window. Updates include episode metadata, transcript and attribution changes, not lease/retry bookkeeping. Show/person metadata has a separate lifecycle. Each item returns the stable Arrays record `id`, the publisher `episode_id`, and RFC3339 `updated_at`. For update polling, pass pagination.cursor unchanged with the same window, and filters. Limit and include_raw may change on retry without restarting. An omitted cursor with has_more=false means the window is drained. Start each new window without the old cursor. Offset is not allowed. This is current-state polling, not a snapshot or deletion/history feed: overlap windows and reconcile delayed writes; intermediate versions may coalesce. `podcast_show_id` comes from `/other/podcast/shows`. `podcast_show_name` takes the name instead and resolves it server-side. The response reports the show name as `podcast_show_name`. `speaker` matches a person in the returned `resolved_speakers`, including host, guest and clip attributions. Registry names and aliases match linked identities; stored resolved names also match directly. Publisher-only labels and attributions excluded from the response do not match. Clips are not guests. `resolved_speakers[*].person_id`, `hosts` and `guests` contain only terminal person-registry UUIDs, matching `/persons`. Historical IDs and merged UUIDs are resolved on read. Unlinked speakers keep their name but omit `person_id`; unresolved references are omitted from `hosts` and `guests`. parameters: - description: Show identifier from /other/podcast/shows in: query name: podcast_show_id schema: type: integer - description: >- Show name, resolved to podcast_show_id. A name matching several shows or none is an error naming them, not an empty page; so is one that disagrees with podcast_show_id in: query name: podcast_show_name schema: type: string - description: >- Exact resolved speaker name or registry alias (case-insensitive), including clip in: query name: speaker schema: type: string - description: Publication day YYYY-MM-DD; same value the response reports as date in: query name: date schema: type: string - description: Also attach the original file (default false) in: query name: include_raw schema: type: boolean - description: Max rows, 1-50 (default 10) in: query name: limit schema: type: integer - description: Rows to skip (default 0) in: query name: offset schema: type: integer - description: >- Inclusive episode update bound (Unix seconds); requires updated_before in: query name: updated_after schema: type: integer format: int64 - description: >- Exclusive episode update bound (Unix seconds); must exceed updated_after in: query name: updated_before schema: type: integer format: int64 - description: Opaque update-window continuation; incompatible with offset in: query name: cursor schema: type: string responses: '200': description: Transcripts content: application/json: schema: allOf: - $ref: '#/components/schemas/model.APIResponse' - properties: data: items: $ref: '#/components/schemas/model.PodcastTranscript' type: array type: object '400': description: Invalid parameters content: application/json: schema: allOf: - $ref: '#/components/schemas/model.APIResponse' - properties: error: $ref: '#/components/schemas/model.APIError' type: object '500': description: Internal server error content: application/json: schema: allOf: - $ref: '#/components/schemas/model.APIResponse' - properties: error: $ref: '#/components/schemas/model.APIError' type: object security: - ApiKeyAuth: [] components: schemas: model.APIResponse: properties: data: {} error: $ref: '#/components/schemas/model.APIError' metadata: {} pagination: {} request_id: type: string success: type: boolean type: object model.PodcastTranscript: properties: created_at: type: string date: type: string episode_id: description: >- EpisodeID identifies the episode, not the show. The publisher's RSS guid, or a key derived from the feed entry when none was given. type: string guests: items: type: string type: array has_timestamps: description: HasTimestamps says whether the turns carry "@ seconds". type: boolean hosts: description: |- Hosts and Guests restate ResolvedSpeakers by role, as person ids for joining against /persons. Only terminal registry UUIDs are returned; unresolved references are omitted. Each id appears once. items: type: string type: array id: description: ID identifies the Arrays record, unlike the publisher's EpisodeID. type: integer podcast_show_id: description: |- PodcastShowID identifies the show, not the episode. Pass it back as podcast_show_id to page or filter without resolving the name again. type: integer podcast_show_name: description: |- PodcastShowName is the show's name, as the podcast_show_name filter matches it. type: string raw_transcript: description: >- RawTranscript is the publisher's file as served, shaped by TranscriptType. Only when include_raw=true. type: string resolved_speakers: additionalProperties: $ref: '#/components/schemas/model.ResolvedSpeaker' description: >- ResolvedSpeakers is who each numbered voice turned out to be. Keys are speaker numbers; join on PersonID, not the key. Entries below 0.4 confidence are dropped server-side. type: object speaker_recognition_status: type: string speaker_recognition_version: type: string speaker_review_reason: type: string speaker_review_required: type: boolean transcript: description: |- Transcript is the normalized body: one turn per line, e.g. "[spk 0 @ 12.5] ...". type: string transcript_source: description: |- TranscriptSource is official or asr. Not the file format; that is TranscriptType. type: string transcript_type: description: >- TranscriptType is the source file's format: vtt, srt, html, plain, json, or deepgram-json. type: string transcript_url: description: >- TranscriptURL is where the document was fetched from. Absent on episodes we transcribed ourselves. type: string updated_at: description: >- UpdatedAt tracks episode metadata, transcript and attribution changes. It is not a content version. Show/person metadata has its own lifecycle. type: string type: object model.APIError: properties: code: example: RESOURCE_NOT_FOUND type: string details: items: $ref: '#/components/schemas/model.APIErrorDetail' type: array docs_url: type: string examples: items: type: string type: array hint: type: string message: example: The requested resource was not found. type: string suggestions: items: type: string type: array type: object model.ResolvedSpeaker: properties: attribution_source: description: >- AttributionSource is how the name was reached: intro_selfstated, intro_introduced, signoff, title, websearch, or adjudicated. Per speaker, since one episode mixes them. type: string confidence: description: >- Confidence runs 0..1. Below 0.4 is already dropped; at or above 0.7 use it, between the two show it as uncertain. A few rows predate the field and report 0. type: number name: description: >- Name is the spelling found on a source, not the one in the transcript: a transcriber hears Lisa Abramowicz as "Lisa Bromwitz". type: string person_id: description: >- PersonID is the terminal person-registry UUID, absent when nothing could be linked. Legacy identifiers and merged UUIDs are resolved before return. type: string presence: description: >- Presence restates Role for callers written against the field that preceded it: archival where Role is clip, evidenced otherwise. Derived, never stored, so it cannot disagree with Role. Empty when Role is, since evidenced would claim the slot took part. enum: - evidenced - archival example: archival type: string role: description: Role is host, guest, or clip. enum: - host - guest - clip example: clip type: string type: object model.APIErrorDetail: properties: field: type: string got: type: string reason: type: string type: object securitySchemes: ApiKeyAuth: description: 'API Key authentication. Example: "your-api-key-here"' in: header name: X-API-Key type: apiKey ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.