@speakai/mcp-server 1.23.0 → 1.24.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (3) hide show
  1. package/README.md +6 -5
  2. package/dist/index.js +126 -258
  3. package/package.json +1 -1
package/README.md CHANGED
@@ -242,7 +242,7 @@ For questions about data handling, see [speakai.co/privacy](https://speakai.co/p
242
242
 
243
243
  ## What you can do once installed
244
244
 
245
- Speak AI ships 113 tools your AI assistant can call. You don't memorize them — Claude/ChatGPT pick the right ones based on what you ask. Examples by category:
245
+ Speak AI ships 114 tools your AI assistant can call. You don't memorize them — Claude/ChatGPT pick the right ones based on what you ask. Examples by category:
246
246
 
247
247
  | Ask | Tools used (auto) |
248
248
  |---|---|
@@ -394,17 +394,18 @@ SPEAK_API_KEY=your-key npx @speakai/mcp-server
394
394
  | `SPEAK_ACCESS_TOKEN` | No | Auto-managed | JWT access token (auto-fetched and refreshed) |
395
395
  | `SPEAK_BASE_URL` | No | `https://api.speakai.co` | API base URL |
396
396
 
397
- ### MCP Tools (113)
397
+ ### MCP Tools (114)
398
398
 
399
399
  <details>
400
- <summary>Media (17 tools)</summary>
400
+ <summary>Media (18 tools)</summary>
401
401
 
402
402
  | Tool | Description |
403
403
  |---|---|
404
404
  | `get_signed_upload_url` | Get a pre-signed S3 URL for direct file upload |
405
- | `upload_media` | Upload media from a URL — a direct/public file URL or a social/video link (YouTube, Instagram, TikTok, X, Facebook, Reddit, SoundCloud, …) resolved automatically. |
405
+ | `upload_media` | Upload media from a URL — a direct/public file URL, a pre-signed S3 URL, or a social/video page link resolved automatically (YouTube, TikTok, Instagram, X/Twitter, Facebook, Reddit, SoundCloud, Twitch, Dailymotion, Streamable, Snapchat, Pinterest, Tumblr, Bilibili, VK, OK.ru, Rutube). Vimeo and Loom page links are not supported. |
406
406
  | `upload_local_file` | Upload a local file directly from disk |
407
- | `upload_and_analyze` | Upload media from a URL (direct file or social/video link, resolved automatically) and return its `media_id` immediately. Poll `get_media_status` until `processed`, then call `get_media_insights` for AI summaries. |
407
+ | `upload_and_analyze` | Upload media from a URL (direct file, or any social/video page link `upload_media` accepts — resolved automatically) and return its `media_id` immediately. Poll `get_media_status` until `processed`, then call `get_media_insights` for AI summaries. Set `mediaType` when the user has said which they want; leave it off otherwise and the server picks the best available track. |
408
+ | `upload_and_analyze_batch` | Upload up to 25 URLs in one call, 5 at a time. Each URL is reported as uploaded or failed with its reason, so one bad link does not sink the batch. Use this instead of calling `upload_and_analyze` in a loop. |
408
409
  | `list_media` | List and search media files with filters, pagination, and optional inline data (transcripts, speakers, keywords) via `include` param |
409
410
  | `get_media_insights` | Get AI insights — topics, sentiment, summaries, action items |
410
411
  | `get_transcript` | Get full transcript with speaker labels and timestamps |
package/dist/index.js CHANGED
@@ -283,7 +283,7 @@ var init_auth = __esm({
283
283
  });
284
284
 
285
285
  // node_modules/@speakai/shared/dist/enums/automation.js
286
- var AutomationTrigger, AutomationAction, AutomationStepType, AutomationRunStatus, AutomationIOType, AutomationRunType, AutomationScheduleTimePeriod, AssistantType;
286
+ var AutomationTrigger, AutomationAction, AutomationRunType, AutomationScheduleTimePeriod, AssistantType;
287
287
  var init_automation = __esm({
288
288
  "node_modules/@speakai/shared/dist/enums/automation.js"() {
289
289
  "use strict";
@@ -291,38 +291,11 @@ var init_automation = __esm({
291
291
  AutomationTrigger2["FOLDERS"] = "folders";
292
292
  AutomationTrigger2["TAGS"] = "tags";
293
293
  AutomationTrigger2["KEYWORDS"] = "keywords";
294
- AutomationTrigger2["COMPOSIO"] = "composio";
295
- AutomationTrigger2["WEBHOOK"] = "webhook";
296
294
  })(AutomationTrigger || (AutomationTrigger = {}));
297
295
  (function(AutomationAction2) {
298
296
  AutomationAction2["MAGIC_PROMPT"] = "magic-prompt";
299
297
  AutomationAction2["TRANSLATION"] = "translation";
300
298
  })(AutomationAction || (AutomationAction = {}));
301
- (function(AutomationStepType2) {
302
- AutomationStepType2["TRIGGER"] = "trigger";
303
- AutomationStepType2["MAGIC_PROMPT"] = "magic-prompt";
304
- AutomationStepType2["TRANSLATION"] = "translation";
305
- AutomationStepType2["COMPOSIO_ACTION"] = "composio-action";
306
- AutomationStepType2["FILTER"] = "filter";
307
- AutomationStepType2["SPEAK_UPLOAD"] = "speak-upload";
308
- AutomationStepType2["NOTIFY"] = "notify";
309
- AutomationStepType2["OUTBOUND_WEBHOOK"] = "outbound-webhook";
310
- AutomationStepType2["CONDITION"] = "condition";
311
- })(AutomationStepType || (AutomationStepType = {}));
312
- (function(AutomationRunStatus2) {
313
- AutomationRunStatus2["PENDING"] = "pending";
314
- AutomationRunStatus2["RUNNING"] = "running";
315
- AutomationRunStatus2["COMPLETED"] = "completed";
316
- AutomationRunStatus2["FAILED"] = "failed";
317
- AutomationRunStatus2["KILLED"] = "killed";
318
- })(AutomationRunStatus || (AutomationRunStatus = {}));
319
- (function(AutomationIOType2) {
320
- AutomationIOType2["FILE"] = "file";
321
- AutomationIOType2["MEDIA"] = "media";
322
- AutomationIOType2["INSIGHT"] = "insight";
323
- AutomationIOType2["NOTIFY"] = "notify";
324
- AutomationIOType2["DATA"] = "data";
325
- })(AutomationIOType || (AutomationIOType = {}));
326
299
  (function(AutomationRunType2) {
327
300
  AutomationRunType2["INSTANT"] = "instant";
328
301
  AutomationRunType2["SCHEDULE"] = "schedule";
@@ -414,7 +387,6 @@ var init_embed = __esm({
414
387
  (function(EmbedType2) {
415
388
  EmbedType2["MEDIA_PLAYER"] = "mediaPlayer";
416
389
  EmbedType2["REPOSITORY"] = "repository";
417
- EmbedType2["DASHBOARD"] = "dashboard";
418
390
  })(EmbedType || (EmbedType = {}));
419
391
  (function(ImageSelectionType2) {
420
392
  ImageSelectionType2["LOGO"] = "logo";
@@ -514,18 +486,6 @@ var init_filter = __esm({
514
486
  }
515
487
  });
516
488
 
517
- // node_modules/@speakai/shared/dist/enums/integration.js
518
- var IntegrationAuthType;
519
- var init_integration = __esm({
520
- "node_modules/@speakai/shared/dist/enums/integration.js"() {
521
- "use strict";
522
- (function(IntegrationAuthType2) {
523
- IntegrationAuthType2["OAUTH"] = "oauth";
524
- IntegrationAuthType2["API_KEY"] = "api_key";
525
- })(IntegrationAuthType || (IntegrationAuthType = {}));
526
- }
527
- });
528
-
529
489
  // node_modules/@speakai/shared/dist/enums/media.js
530
490
  var MediaType, MediaState, MediaPrivacyMode, MediaInsightType, MediaInsightStatus, MediaProcessType;
531
491
  var init_media = __esm({
@@ -1003,72 +963,10 @@ var init_webhook = __esm({
1003
963
  (function(WebhookEventSource2) {
1004
964
  WebhookEventSource2["SPEAK"] = "speak";
1005
965
  WebhookEventSource2["ZAPIER"] = "zapier";
1006
- WebhookEventSource2["N8N"] = "n8n";
1007
- WebhookEventSource2["PIPEDREAM"] = "pipedream";
1008
- WebhookEventSource2["MAKE"] = "make";
1009
966
  })(WebhookEventSource || (WebhookEventSource = {}));
1010
967
  }
1011
968
  });
1012
969
 
1013
- // node_modules/@speakai/shared/dist/enums/llm.js
1014
- var LLMProvider, LLMModels;
1015
- var init_llm = __esm({
1016
- "node_modules/@speakai/shared/dist/enums/llm.js"() {
1017
- "use strict";
1018
- (function(LLMProvider2) {
1019
- LLMProvider2["OPENAI"] = "openai";
1020
- LLMProvider2["GOOGLE"] = "google";
1021
- LLMProvider2["ANTHROPIC"] = "anthropic";
1022
- LLMProvider2["OPENROUTER"] = "openrouter";
1023
- })(LLMProvider || (LLMProvider = {}));
1024
- (function(LLMModels2) {
1025
- LLMModels2["GPT_3_5"] = "gpt-3.5";
1026
- LLMModels2["GPT_3_5_TURBO_16K"] = "gpt-3.5-turbo-16k";
1027
- LLMModels2["GPT_3_5_TURBO_0125"] = "gpt-3.5-turbo-0125";
1028
- LLMModels2["GPT_4"] = "gpt-4";
1029
- LLMModels2["GPT_4_1106_PREVIEW"] = "gpt-4-1106-preview";
1030
- LLMModels2["GPT_4_TURBO"] = "gpt-4-turbo";
1031
- LLMModels2["GPT_4_O_2024_05_13"] = "gpt-4o-2024-05-13";
1032
- LLMModels2["GPT_4O"] = "gpt-4o";
1033
- LLMModels2["GPT_4O_MINI"] = "gpt-4o-mini";
1034
- LLMModels2["GPT_4_O_2024_08_06"] = "gpt-4o-2024-08-06";
1035
- LLMModels2["GPT_4_MINI_2024_07_18"] = "gpt-4o-mini-2024-07-18";
1036
- LLMModels2["GPT_4_1_2025_04_14"] = "gpt-4.1-2025-04-14";
1037
- LLMModels2["GPT_5_1_2025_11_13"] = "gpt-5.1-2025-11-13";
1038
- LLMModels2["GPT_5_2"] = "gpt-5.2";
1039
- LLMModels2["GPT_5_4"] = "gpt-5.4";
1040
- LLMModels2["GPT_5_4_MINI"] = "gpt-5.4-mini";
1041
- LLMModels2["GPT_5_4_MINI_2026_03_17"] = "gpt-5.4-mini-2026-03-17";
1042
- LLMModels2["GPT_5_4_NANO"] = "gpt-5.4-nano";
1043
- LLMModels2["GPT_5_5"] = "gpt-5.5";
1044
- LLMModels2["GPT_5_5_THINKING"] = "gpt-5.5-thinking";
1045
- LLMModels2["GPT_5_6_SOL"] = "gpt-5.6-sol";
1046
- LLMModels2["GPT_5_6_TERRA"] = "gpt-5.6-terra";
1047
- LLMModels2["GPT_5_6_LUNA"] = "gpt-5.6-luna";
1048
- LLMModels2["CLAUDE_2"] = "claude-2";
1049
- LLMModels2["CLAUDE_3_5_SONNET"] = "claude-3-5-sonnet";
1050
- LLMModels2["CLAUDE_3_5_SONNET_20241022"] = "claude-3-5-sonnet-20241022";
1051
- LLMModels2["CLAUDE_3_7_SONNET_LATEST"] = "claude-3-7-sonnet-latest";
1052
- LLMModels2["CLAUDE_HAIKU_4_5"] = "claude-haiku-4-5";
1053
- LLMModels2["CLAUDE_SONNET_4_6"] = "claude-sonnet-4-6";
1054
- LLMModels2["CLAUDE_SONNET_5"] = "claude-sonnet-5";
1055
- LLMModels2["CLAUDE_OPUS_4_8"] = "claude-opus-4-8";
1056
- LLMModels2["GEMINI_1_5_PRO"] = "gemini-1.5-pro";
1057
- LLMModels2["GEMINI_1_5_FLASH"] = "gemini-1.5-flash";
1058
- LLMModels2["GEMINI_2_0_FLASH"] = "gemini-2.0-flash";
1059
- LLMModels2["GEMINI_2_5_PRO"] = "gemini-2.5-pro";
1060
- LLMModels2["GEMINI_2_5_FLASH"] = "gemini-2.5-flash";
1061
- LLMModels2["GEMINI_2_5_FLASH_LITE"] = "gemini-2.5-flash-lite";
1062
- LLMModels2["GEMINI_3_FLASH_PREVIEW"] = "gemini-3-flash-preview";
1063
- LLMModels2["GEMINI_3_1_FLASH_LITE"] = "gemini-3.1-flash-lite";
1064
- LLMModels2["GEMINI_3_1_PRO_PREVIEW"] = "gemini-3.1-pro-preview";
1065
- LLMModels2["GEMINI_3_5_FLASH"] = "gemini-3.5-flash";
1066
- LLMModels2["GROK_4_5"] = "x-ai/grok-4.5";
1067
- LLMModels2["GLM_5_2"] = "z-ai/glm-5.2";
1068
- })(LLMModels || (LLMModels = {}));
1069
- }
1070
- });
1071
-
1072
970
  // node_modules/@speakai/shared/dist/enums/index.js
1073
971
  var init_enums = __esm({
1074
972
  "node_modules/@speakai/shared/dist/enums/index.js"() {
@@ -1083,7 +981,6 @@ var init_enums = __esm({
1083
981
  init_export();
1084
982
  init_fields();
1085
983
  init_filter();
1086
- init_integration();
1087
984
  init_media();
1088
985
  init_meeting();
1089
986
  init_notification();
@@ -1096,7 +993,6 @@ var init_enums = __esm({
1096
993
  init_translation();
1097
994
  init_user();
1098
995
  init_webhook();
1099
- init_llm();
1100
996
  }
1101
997
  });
1102
998
 
@@ -1135,13 +1031,6 @@ var init_folder = __esm({
1135
1031
  }
1136
1032
  });
1137
1033
 
1138
- // node_modules/@speakai/shared/dist/interfaces/integration.js
1139
- var init_integration2 = __esm({
1140
- "node_modules/@speakai/shared/dist/interfaces/integration.js"() {
1141
- "use strict";
1142
- }
1143
- });
1144
-
1145
1034
  // node_modules/@speakai/shared/dist/interfaces/recorder.js
1146
1035
  var init_recorder2 = __esm({
1147
1036
  "node_modules/@speakai/shared/dist/interfaces/recorder.js"() {
@@ -1233,21 +1122,6 @@ var init_clip2 = __esm({
1233
1122
  }
1234
1123
  });
1235
1124
 
1236
- // node_modules/@speakai/shared/dist/utils/dashboard-spec.js
1237
- var init_dashboard_spec = __esm({
1238
- "node_modules/@speakai/shared/dist/utils/dashboard-spec.js"() {
1239
- "use strict";
1240
- }
1241
- });
1242
-
1243
- // node_modules/@speakai/shared/dist/interfaces/dashboard.js
1244
- var init_dashboard = __esm({
1245
- "node_modules/@speakai/shared/dist/interfaces/dashboard.js"() {
1246
- "use strict";
1247
- init_dashboard_spec();
1248
- }
1249
- });
1250
-
1251
1125
  // node_modules/@speakai/shared/dist/interfaces/index.js
1252
1126
  var init_interfaces = __esm({
1253
1127
  "node_modules/@speakai/shared/dist/interfaces/index.js"() {
@@ -1257,7 +1131,6 @@ var init_interfaces = __esm({
1257
1131
  init_transcript();
1258
1132
  init_text();
1259
1133
  init_folder();
1260
- init_integration2();
1261
1134
  init_recorder2();
1262
1135
  init_embed2();
1263
1136
  init_automation2();
@@ -1271,7 +1144,6 @@ var init_interfaces = __esm({
1271
1144
  init_calendar2();
1272
1145
  init_category();
1273
1146
  init_clip2();
1274
- init_dashboard();
1275
1147
  }
1276
1148
  });
1277
1149
 
@@ -1282,70 +1154,6 @@ var init_transcript2 = __esm({
1282
1154
  }
1283
1155
  });
1284
1156
 
1285
- // node_modules/@speakai/shared/dist/pricing/modelPricing.js
1286
- var MODEL_PRICING;
1287
- var init_modelPricing = __esm({
1288
- "node_modules/@speakai/shared/dist/pricing/modelPricing.js"() {
1289
- "use strict";
1290
- init_llm();
1291
- MODEL_PRICING = {
1292
- // ═══════════════ OpenAI ═══════════════
1293
- // Deprecated
1294
- [LLMModels.GPT_3_5]: { inputPerMillion: 0.5, outputPerMillion: 1.5, provider: LLMProvider.OPENAI },
1295
- [LLMModels.GPT_3_5_TURBO_16K]: { inputPerMillion: 3, outputPerMillion: 4, provider: LLMProvider.OPENAI },
1296
- [LLMModels.GPT_3_5_TURBO_0125]: { inputPerMillion: 0.5, outputPerMillion: 1.5, provider: LLMProvider.OPENAI },
1297
- [LLMModels.GPT_4]: { inputPerMillion: 30, outputPerMillion: 60, provider: LLMProvider.OPENAI },
1298
- [LLMModels.GPT_4_1106_PREVIEW]: { inputPerMillion: 10, outputPerMillion: 30, provider: LLMProvider.OPENAI },
1299
- [LLMModels.GPT_4_TURBO]: { inputPerMillion: 10, outputPerMillion: 30, provider: LLMProvider.OPENAI },
1300
- [LLMModels.GPT_4_O_2024_05_13]: { inputPerMillion: 5, outputPerMillion: 15, provider: LLMProvider.OPENAI },
1301
- // Live
1302
- [LLMModels.GPT_4O]: { inputPerMillion: 2.5, outputPerMillion: 10, cachedInputPerMillion: 1.25, provider: LLMProvider.OPENAI },
1303
- [LLMModels.GPT_4O_MINI]: { inputPerMillion: 0.15, outputPerMillion: 0.6, cachedInputPerMillion: 0.075, provider: LLMProvider.OPENAI },
1304
- [LLMModels.GPT_4_O_2024_08_06]: { inputPerMillion: 2.5, outputPerMillion: 10, cachedInputPerMillion: 1.25, provider: LLMProvider.OPENAI },
1305
- [LLMModels.GPT_4_MINI_2024_07_18]: { inputPerMillion: 0.15, outputPerMillion: 0.6, cachedInputPerMillion: 0.075, provider: LLMProvider.OPENAI },
1306
- [LLMModels.GPT_4_1_2025_04_14]: { inputPerMillion: 2, outputPerMillion: 8, cachedInputPerMillion: 0.5, provider: LLMProvider.OPENAI },
1307
- [LLMModels.GPT_5_1_2025_11_13]: { inputPerMillion: 1.25, outputPerMillion: 10, cachedInputPerMillion: 0.125, provider: LLMProvider.OPENAI },
1308
- [LLMModels.GPT_5_2]: { inputPerMillion: 1.75, outputPerMillion: 14, cachedInputPerMillion: 0.175, provider: LLMProvider.OPENAI },
1309
- [LLMModels.GPT_5_4]: { inputPerMillion: 2.5, outputPerMillion: 15, cachedInputPerMillion: 0.25, longContextThresholdTokens: 272e3, inputPerMillionLong: 5, outputPerMillionLong: 22.5, cachedInputPerMillionLong: 0.5, provider: LLMProvider.OPENAI },
1310
- [LLMModels.GPT_5_4_MINI]: { inputPerMillion: 0.75, outputPerMillion: 4.5, cachedInputPerMillion: 0.075, provider: LLMProvider.OPENAI },
1311
- [LLMModels.GPT_5_4_MINI_2026_03_17]: { inputPerMillion: 0.75, outputPerMillion: 4.5, cachedInputPerMillion: 0.075, provider: LLMProvider.OPENAI },
1312
- [LLMModels.GPT_5_4_NANO]: { inputPerMillion: 0.2, outputPerMillion: 1.25, cachedInputPerMillion: 0.02, provider: LLMProvider.OPENAI },
1313
- [LLMModels.GPT_5_5]: { inputPerMillion: 5, outputPerMillion: 30, cachedInputPerMillion: 0.5, longContextThresholdTokens: 272e3, inputPerMillionLong: 10, outputPerMillionLong: 45, cachedInputPerMillionLong: 1, provider: LLMProvider.OPENAI },
1314
- [LLMModels.GPT_5_5_THINKING]: { inputPerMillion: 5, outputPerMillion: 30, cachedInputPerMillion: 0.5, longContextThresholdTokens: 272e3, inputPerMillionLong: 10, outputPerMillionLong: 45, cachedInputPerMillionLong: 1, provider: LLMProvider.OPENAI },
1315
- // gpt-5.5 reasoning mode (same rate)
1316
- [LLMModels.GPT_5_6_SOL]: { inputPerMillion: 5, outputPerMillion: 30, cachedInputPerMillion: 0.5, provider: LLMProvider.OPENAI },
1317
- [LLMModels.GPT_5_6_TERRA]: { inputPerMillion: 2.5, outputPerMillion: 15, cachedInputPerMillion: 0.25, provider: LLMProvider.OPENAI },
1318
- [LLMModels.GPT_5_6_LUNA]: { inputPerMillion: 1, outputPerMillion: 6, cachedInputPerMillion: 0.1, provider: LLMProvider.OPENAI },
1319
- // ═══════════════ Anthropic (cache read = 0.1x input) ═══════════════
1320
- // Deprecated
1321
- [LLMModels.CLAUDE_2]: { inputPerMillion: 8, outputPerMillion: 24, provider: LLMProvider.ANTHROPIC },
1322
- [LLMModels.CLAUDE_3_5_SONNET]: { inputPerMillion: 3, outputPerMillion: 15, cachedInputPerMillion: 0.3, provider: LLMProvider.ANTHROPIC },
1323
- [LLMModels.CLAUDE_3_5_SONNET_20241022]: { inputPerMillion: 3, outputPerMillion: 15, cachedInputPerMillion: 0.3, provider: LLMProvider.ANTHROPIC },
1324
- [LLMModels.CLAUDE_3_7_SONNET_LATEST]: { inputPerMillion: 3, outputPerMillion: 15, cachedInputPerMillion: 0.3, provider: LLMProvider.ANTHROPIC },
1325
- // Live
1326
- [LLMModels.CLAUDE_HAIKU_4_5]: { inputPerMillion: 1, outputPerMillion: 5, cachedInputPerMillion: 0.1, provider: LLMProvider.ANTHROPIC },
1327
- [LLMModels.CLAUDE_SONNET_4_6]: { inputPerMillion: 3, outputPerMillion: 15, cachedInputPerMillion: 0.3, provider: LLMProvider.ANTHROPIC },
1328
- [LLMModels.CLAUDE_SONNET_5]: { inputPerMillion: 3, outputPerMillion: 15, cachedInputPerMillion: 0.3, provider: LLMProvider.ANTHROPIC },
1329
- [LLMModels.CLAUDE_OPUS_4_8]: { inputPerMillion: 5, outputPerMillion: 25, cachedInputPerMillion: 0.5, provider: LLMProvider.ANTHROPIC },
1330
- // ═══════════════ Google Gemini ═══════════════
1331
- // Deprecated
1332
- [LLMModels.GEMINI_1_5_PRO]: { inputPerMillion: 1.25, outputPerMillion: 5, provider: LLMProvider.GOOGLE },
1333
- [LLMModels.GEMINI_1_5_FLASH]: { inputPerMillion: 0.075, outputPerMillion: 0.3, cachedInputPerMillion: 0.01875, provider: LLMProvider.GOOGLE },
1334
- [LLMModels.GEMINI_2_0_FLASH]: { inputPerMillion: 0.1, outputPerMillion: 0.4, cachedInputPerMillion: 0.025, provider: LLMProvider.GOOGLE },
1335
- // Live
1336
- [LLMModels.GEMINI_2_5_PRO]: { inputPerMillion: 1.25, outputPerMillion: 10, longContextThresholdTokens: 2e5, inputPerMillionLong: 2.5, outputPerMillionLong: 15, provider: LLMProvider.GOOGLE },
1337
- [LLMModels.GEMINI_2_5_FLASH]: { inputPerMillion: 0.3, outputPerMillion: 2.5, cachedInputPerMillion: 0.03, provider: LLMProvider.GOOGLE },
1338
- [LLMModels.GEMINI_2_5_FLASH_LITE]: { inputPerMillion: 0.1, outputPerMillion: 0.4, cachedInputPerMillion: 0.01, provider: LLMProvider.GOOGLE },
1339
- [LLMModels.GEMINI_3_FLASH_PREVIEW]: { inputPerMillion: 0.5, outputPerMillion: 3, cachedInputPerMillion: 0.05, provider: LLMProvider.GOOGLE },
1340
- [LLMModels.GEMINI_3_1_FLASH_LITE]: { inputPerMillion: 0.25, outputPerMillion: 1.5, provider: LLMProvider.GOOGLE },
1341
- [LLMModels.GEMINI_3_1_PRO_PREVIEW]: { inputPerMillion: 2, outputPerMillion: 12, longContextThresholdTokens: 2e5, inputPerMillionLong: 4, outputPerMillionLong: 18, provider: LLMProvider.GOOGLE },
1342
- [LLMModels.GEMINI_3_5_FLASH]: { inputPerMillion: 1.5, outputPerMillion: 9, cachedInputPerMillion: 0.15, provider: LLMProvider.GOOGLE },
1343
- [LLMModels.GROK_4_5]: { inputPerMillion: 2.2, outputPerMillion: 6.6, cachedInputPerMillion: 0.22, provider: LLMProvider.OPENROUTER },
1344
- [LLMModels.GLM_5_2]: { inputPerMillion: 1.023, outputPerMillion: 3.3, cachedInputPerMillion: 0.1023, provider: LLMProvider.OPENROUTER }
1345
- };
1346
- }
1347
- });
1348
-
1349
1157
  // node_modules/@speakai/shared/dist/index.js
1350
1158
  var init_dist = __esm({
1351
1159
  "node_modules/@speakai/shared/dist/index.js"() {
@@ -1353,8 +1161,6 @@ var init_dist = __esm({
1353
1161
  init_enums();
1354
1162
  init_interfaces();
1355
1163
  init_transcript2();
1356
- init_dashboard_spec();
1357
- init_modelPricing();
1358
1164
  }
1359
1165
  });
1360
1166
 
@@ -1372,7 +1178,7 @@ function getMimeType(filePath) {
1372
1178
  function detectMediaType(filePath) {
1373
1179
  return isVideoFile(filePath) ? "video" : "audio";
1374
1180
  }
1375
- var path, VIDEO_EXTENSIONS, MIME_TYPES;
1181
+ var path, VIDEO_EXTENSIONS, MIME_TYPES, SUPPORTED_URL_SOURCES, UNSUPPORTED_URL_SOURCES;
1376
1182
  var init_media_utils = __esm({
1377
1183
  "src/media-utils.ts"() {
1378
1184
  "use strict";
@@ -1389,6 +1195,8 @@ var init_media_utils = __esm({
1389
1195
  ".mkv": "video/x-matroska",
1390
1196
  ".wmv": "video/x-ms-wmv"
1391
1197
  };
1198
+ SUPPORTED_URL_SOURCES = "YouTube, TikTok, Instagram, X/Twitter, Facebook, Reddit, SoundCloud, Twitch, Dailymotion, Streamable, Snapchat, Pinterest, Tumblr, Bilibili, VK, OK.ru and Rutube";
1199
+ UNSUPPORTED_URL_SOURCES = "Vimeo and Loom page links are not supported.";
1392
1200
  }
1393
1201
  });
1394
1202
 
@@ -1407,7 +1215,7 @@ function register(server, client) {
1407
1215
  "Get a pre-signed S3 URL for direct file upload to Speak AI storage. After getting the URL, PUT your file to it, then call upload_media with the S3 URL. For a simpler workflow, use upload_local_file instead which handles all steps automatically.",
1408
1216
  {
1409
1217
  filename: import_zod2.z.string().min(1).describe('Original filename including extension, e.g. "interview.mp4". The extension decides audio vs video and the storage path, so it must match the real file \u2014 a video named ".mp3" is stored as audio and can never be analysed as video.'),
1410
- isVideo: import_zod2.z.boolean().optional().describe("Override the audio/video decision. Omit to derive it from the filename extension (recommended); only set this when the extension does not reflect the real container."),
1218
+ mediaType: import_zod2.z.enum([MediaType.AUDIO, MediaType.VIDEO]).optional().describe('Type of media: "audio" or "video". It decides the storage path, so it has to match the real file. Send it whenever the user has told you which they want, or when the filename extension does not reflect the real container. Omit it to derive the type from the extension, which is the only evidence available here because the bytes do not exist yet.'),
1411
1219
  mimeType: import_zod2.z.string().optional().describe('MIME type, e.g. "video/mp4". Omit to derive it from the filename extension.')
1412
1220
  },
1413
1221
  {
@@ -1417,12 +1225,12 @@ function register(server, client) {
1417
1225
  idempotentHint: false,
1418
1226
  openWorldHint: false
1419
1227
  },
1420
- async ({ isVideo, filename, mimeType }) => {
1228
+ async ({ mediaType, filename, mimeType }) => {
1421
1229
  try {
1422
- const resolvedIsVideo = isVideo ?? isVideoFile(filename);
1230
+ const resolvedMediaType = mediaType ?? (isVideoFile(filename) ? MediaType.VIDEO : MediaType.AUDIO);
1423
1231
  const resolvedMimeType = mimeType ?? getMimeType(filename);
1424
1232
  const result = await api.get("/v1/media/upload/signedurl", {
1425
- params: { isVideo: resolvedIsVideo, filename, mimeType: resolvedMimeType }
1233
+ params: { mediaType: resolvedMediaType, filename, mimeType: resolvedMimeType }
1426
1234
  });
1427
1235
  return {
1428
1236
  content: [
@@ -1440,11 +1248,11 @@ function register(server, client) {
1440
1248
  registerSpeakTool(
1441
1249
  server,
1442
1250
  "upload_media",
1443
- "Upload media from a URL \u2014 a direct/public file URL, a pre-signed S3 URL, or a shareable social/video link (YouTube, Instagram, TikTok, X, Facebook, Reddit, SoundCloud, and similar) which Speak resolves to the underlying media automatically. Processing is asynchronous \u2014 after uploading, use get_media_status to poll until state is 'processed' (typically 1-3 minutes for audio under 60 min), then use get_transcript and get_media_insights to retrieve results. For a single call that handles everything, use upload_and_analyze instead. For local files, use upload_local_file. (Vimeo links are not yet supported.)",
1251
+ `Upload media from a URL \u2014 a direct/public file URL, a pre-signed S3 URL, or a shareable social/video page link, which Speak resolves to the underlying media automatically. Supported page links: ${SUPPORTED_URL_SOURCES}. ${UNSUPPORTED_URL_SOURCES} Processing is asynchronous \u2014 after uploading, use get_media_status to poll until state is 'processed' (typically 1-3 minutes for audio under 60 min), then use get_transcript and get_media_insights to retrieve results. For a single call that handles everything, use upload_and_analyze instead. For local files, use upload_local_file.`,
1444
1252
  {
1445
1253
  name: import_zod2.z.string().min(1).describe("Display name for the media file"),
1446
- url: import_zod2.z.string().describe("Direct/public media file URL, pre-signed S3 URL, or a shareable social/video page link (e.g. an Instagram reel or TikTok URL) \u2014 page links are resolved to the underlying media server-side."),
1447
- mediaType: import_zod2.z.enum([MediaType.AUDIO, MediaType.VIDEO]).optional().describe(`Type of media: "audio" or "video". Omit to derive it from the URL's file extension (recommended). Set it only when the URL has no usable extension; a video labelled "audio" here can never be analysed as video.`),
1254
+ url: import_zod2.z.string().describe("Direct/public media file URL, pre-signed S3 URL, or a shareable social/video page link \u2014 page links are resolved to the underlying media server-side. See this tool's description for the platforms accepted. Pass the URL the user gave you as-is; do not try to convert it to a file URL first."),
1255
+ mediaType: import_zod2.z.enum([MediaType.AUDIO, MediaType.VIDEO]).optional().describe('Type of media: "audio" or "video". Send it whenever the user has told you which they want \u2014 if they called it an audio file, or asked for audio only, pass "audio"; if they called it a video, pass "video". Otherwise omit it and the server decides: it inspects the actual file for a direct URL, and picks the best track the platform offers for a page link. Do not guess from the URL, because sending a value stops the server inspecting the file, and a video imported as "audio" can never be analysed as video afterwards.'),
1448
1256
  description: import_zod2.z.string().optional().describe("Description of the media file"),
1449
1257
  sourceLanguage: import_zod2.z.string().optional().describe('BCP-47 language code for transcription, e.g. "en-US" or "he-IL"'),
1450
1258
  tags: import_zod2.z.string().optional().describe("Comma-separated tags for the media"),
@@ -1466,9 +1274,7 @@ function register(server, client) {
1466
1274
  },
1467
1275
  async (body) => {
1468
1276
  try {
1469
- const urlPath = body.url.split(/[?#]/)[0];
1470
- const mediaType = body.mediaType ?? detectMediaType(urlPath);
1471
- const result = await api.post("/v1/media/upload", { ...body, mediaType });
1277
+ const result = await api.post("/v1/media/upload", body);
1472
1278
  return {
1473
1279
  content: [
1474
1280
  { type: "text", text: JSON.stringify(result.data, null, 2) }
@@ -2052,7 +1858,7 @@ function register2(server, client) {
2052
1858
  readOnlyHint: false,
2053
1859
  destructiveHint: false,
2054
1860
  idempotentHint: false,
2055
- openWorldHint: true
1861
+ openWorldHint: false
2056
1862
  },
2057
1863
  async (body) => {
2058
1864
  try {
@@ -2672,27 +2478,6 @@ var recorder_exports = {};
2672
2478
  __export(recorder_exports, {
2673
2479
  register: () => register5
2674
2480
  });
2675
- function optionsFromMetaType(metaType) {
2676
- const bool = (v, fallback) => typeof v === "boolean" ? v : fallback;
2677
- const upload = metaType?.upload ?? {};
2678
- return {
2679
- audio: bool(metaType?.audio, true),
2680
- video: bool(metaType?.video, true),
2681
- screenShare: bool(metaType?.screenShare, false),
2682
- upload: {
2683
- file: bool(upload.file, true),
2684
- multiple: bool(upload.multiple, false),
2685
- url: bool(upload.url, false)
2686
- },
2687
- liveTranscription: bool(metaType?.liveTranscription, false)
2688
- };
2689
- }
2690
- function addRecorderOptions(recorder) {
2691
- if (recorder && typeof recorder === "object") {
2692
- recorder.options = optionsFromMetaType(recorder.meta?.type);
2693
- }
2694
- return recorder;
2695
- }
2696
2481
  function register5(server, client) {
2697
2482
  const api = client ?? speakClient;
2698
2483
  registerSpeakTool(
@@ -2739,12 +2524,11 @@ function register5(server, client) {
2739
2524
  readOnlyHint: false,
2740
2525
  destructiveHint: false,
2741
2526
  idempotentHint: false,
2742
- openWorldHint: true
2527
+ openWorldHint: false
2743
2528
  },
2744
2529
  async (body) => {
2745
2530
  try {
2746
2531
  const result = await api.post("/v1/recorder/create", body);
2747
- addRecorderOptions(result.data?.data?.recorderData);
2748
2532
  return {
2749
2533
  content: [{ type: "text", text: JSON.stringify(result.data, null, 2) }]
2750
2534
  };
@@ -2775,7 +2559,6 @@ function register5(server, client) {
2775
2559
  async (params) => {
2776
2560
  try {
2777
2561
  const result = await api.get("/v1/recorder", { params });
2778
- result.data?.data?.recorderList?.forEach?.(addRecorderOptions);
2779
2562
  return {
2780
2563
  content: [{ type: "text", text: JSON.stringify(result.data, null, 2) }]
2781
2564
  };
@@ -2802,7 +2585,7 @@ function register5(server, client) {
2802
2585
  readOnlyHint: false,
2803
2586
  destructiveHint: false,
2804
2587
  idempotentHint: false,
2805
- openWorldHint: true
2588
+ openWorldHint: false
2806
2589
  },
2807
2590
  async (body) => {
2808
2591
  try {
@@ -2835,7 +2618,6 @@ function register5(server, client) {
2835
2618
  async ({ recorderId }) => {
2836
2619
  try {
2837
2620
  const result = await api.get(`/v1/recorder/${recorderId}`);
2838
- addRecorderOptions(result.data?.data);
2839
2621
  return {
2840
2622
  content: [{ type: "text", text: JSON.stringify(result.data, null, 2) }]
2841
2623
  };
@@ -2917,12 +2699,11 @@ function register5(server, client) {
2917
2699
  readOnlyHint: false,
2918
2700
  destructiveHint: false,
2919
2701
  idempotentHint: true,
2920
- openWorldHint: true
2702
+ openWorldHint: false
2921
2703
  },
2922
2704
  async ({ recorderId, ...body }) => {
2923
2705
  try {
2924
2706
  const result = await api.put(`/v1/recorder/settings/${recorderId}`, body);
2925
- addRecorderOptions(result.data?.data?.recorderData);
2926
2707
  return {
2927
2708
  content: [{ type: "text", text: JSON.stringify(result.data, null, 2) }]
2928
2709
  };
@@ -2954,7 +2735,7 @@ function register5(server, client) {
2954
2735
  readOnlyHint: false,
2955
2736
  destructiveHint: true,
2956
2737
  idempotentHint: true,
2957
- openWorldHint: true
2738
+ openWorldHint: false
2958
2739
  },
2959
2740
  async ({ recorderId, ...body }) => {
2960
2741
  try {
@@ -2982,7 +2763,7 @@ function register5(server, client) {
2982
2763
  readOnlyHint: false,
2983
2764
  destructiveHint: true,
2984
2765
  idempotentHint: true,
2985
- openWorldHint: true
2766
+ openWorldHint: false
2986
2767
  },
2987
2768
  async ({ recorderId }) => {
2988
2769
  try {
@@ -3055,7 +2836,7 @@ function register6(server, client) {
3055
2836
  readOnlyHint: false,
3056
2837
  destructiveHint: false,
3057
2838
  idempotentHint: false,
3058
- openWorldHint: true
2839
+ openWorldHint: false
3059
2840
  },
3060
2841
  async (body) => {
3061
2842
  try {
@@ -3090,7 +2871,7 @@ function register6(server, client) {
3090
2871
  readOnlyHint: false,
3091
2872
  destructiveHint: false,
3092
2873
  idempotentHint: true,
3093
- openWorldHint: true
2874
+ openWorldHint: false
3094
2875
  },
3095
2876
  async ({ embedId, ...body }) => {
3096
2877
  try {
@@ -4293,7 +4074,7 @@ function register10(server, client) {
4293
4074
  readOnlyHint: false,
4294
4075
  destructiveHint: false,
4295
4076
  idempotentHint: false,
4296
- openWorldHint: true
4077
+ openWorldHint: false
4297
4078
  },
4298
4079
  async (body) => {
4299
4080
  try {
@@ -4329,7 +4110,7 @@ function register10(server, client) {
4329
4110
  readOnlyHint: false,
4330
4111
  destructiveHint: true,
4331
4112
  idempotentHint: true,
4332
- openWorldHint: true
4113
+ openWorldHint: false
4333
4114
  },
4334
4115
  async ({ automationId, ...body }) => {
4335
4116
  try {
@@ -4363,7 +4144,7 @@ function register10(server, client) {
4363
4144
  readOnlyHint: false,
4364
4145
  destructiveHint: false,
4365
4146
  idempotentHint: false,
4366
- openWorldHint: true
4147
+ openWorldHint: false
4367
4148
  },
4368
4149
  async ({ automationId }) => {
4369
4150
  try {
@@ -4392,7 +4173,7 @@ function register10(server, client) {
4392
4173
  readOnlyHint: false,
4393
4174
  destructiveHint: false,
4394
4175
  idempotentHint: true,
4395
- openWorldHint: true
4176
+ openWorldHint: false
4396
4177
  },
4397
4178
  async (body) => {
4398
4179
  try {
@@ -4421,7 +4202,7 @@ function register10(server, client) {
4421
4202
  readOnlyHint: false,
4422
4203
  destructiveHint: true,
4423
4204
  idempotentHint: true,
4424
- openWorldHint: true
4205
+ openWorldHint: false
4425
4206
  },
4426
4207
  async (body) => {
4427
4208
  try {
@@ -5161,7 +4942,7 @@ function register14(server, client) {
5161
4942
  readOnlyHint: false,
5162
4943
  destructiveHint: false,
5163
4944
  idempotentHint: false,
5164
- openWorldHint: true
4945
+ openWorldHint: false
5165
4946
  },
5166
4947
  async (args2) => {
5167
4948
  const { name, trigger, steps, automationId, description, isActive, orTriggers } = args2;
@@ -5410,11 +5191,13 @@ Likely cause: this server rejects filter rules on webhook payload fields (${data
5410
5191
  registerSpeakTool(
5411
5192
  server,
5412
5193
  "upload_and_analyze",
5413
- "Upload and transcribe media from a URL \u2014 a direct/public file URL, OR a shareable social/video link (YouTube, Instagram, TikTok, X, Facebook, Reddit, SoundCloud, and similar), which Speak resolves to the underlying media automatically. Returns media_id immediately; after this returns, poll get_media_status until state is 'processed' (typically 1-3 min for under 60min audio), then call get_media_insights for AI summaries. This async pattern is required for remote MCP transports \u2014 long blocking calls die at proxy idle timeouts. (Vimeo links are not yet supported.)",
5194
+ `Upload and transcribe media from a URL \u2014 a direct/public file URL, OR a shareable social/video page link, which Speak resolves to the underlying media automatically. Supported page links: ${SUPPORTED_URL_SOURCES}. ${UNSUPPORTED_URL_SOURCES} Returns media_id immediately; after this returns, poll get_media_status until state is 'processed' (typically 1-3 min for under 60min audio), then call get_media_insights for AI summaries. This async pattern is required for remote MCP transports \u2014 long blocking calls die at proxy idle timeouts.`,
5414
5195
  {
5415
- url: import_zod15.z.string().describe("Direct/public media file URL, or a shareable social/video page link (e.g. an Instagram reel, TikTok, YouTube, or X post URL) \u2014 page links are resolved to the underlying media server-side. Pass the URL the user gave you as-is."),
5196
+ // A plain literal, not a template: the docs generator drops a tool's whole parameter
5197
+ // table when a description interpolates a value it cannot resolve statically.
5198
+ url: import_zod15.z.string().describe("Direct/public media file URL, or a shareable social/video page link \u2014 page links are resolved to the underlying media server-side. See this tool's description for the platforms accepted. Pass the URL the user gave you as-is; do not try to convert it to a file URL first."),
5416
5199
  name: import_zod15.z.string().optional().describe("Display name for the media (defaults to filename from URL)"),
5417
- mediaType: import_zod15.z.enum([MediaType.AUDIO, MediaType.VIDEO]).optional().describe("Media type (default: audio)"),
5200
+ mediaType: import_zod15.z.enum([MediaType.AUDIO, MediaType.VIDEO]).optional().describe('Type of media: "audio" or "video". Send it whenever the user has told you which they want \u2014 if they called it an audio file, or asked for audio only, pass "audio"; if they called it a video, pass "video". Otherwise omit it and the server decides: it inspects the actual file for a direct URL, and picks the best track the platform offers for a page link. Do not guess from the URL, because sending a value stops the server inspecting the file, and a video imported as "audio" can never be analysed as video afterwards.'),
5418
5201
  sourceLanguage: import_zod15.z.string().optional().describe("BCP-47 language code (e.g., 'en-US', 'he-IL')"),
5419
5202
  folderId: import_zod15.z.string().optional().describe("Folder ID to place the media in"),
5420
5203
  tags: import_zod15.z.string().optional().describe("Comma-separated tags")
@@ -5430,9 +5213,9 @@ Likely cause: this server rejects filter rules on webhook payload fields (${data
5430
5213
  try {
5431
5214
  const uploadBody = {
5432
5215
  name: params.name ?? params.url.split("/").pop()?.split("?")[0] ?? "Upload",
5433
- url: params.url,
5434
- mediaType: params.mediaType ?? "audio"
5216
+ url: params.url
5435
5217
  };
5218
+ if (params.mediaType) uploadBody.mediaType = params.mediaType;
5436
5219
  if (params.sourceLanguage) uploadBody.sourceLanguage = params.sourceLanguage;
5437
5220
  if (params.folderId) uploadBody.folderId = params.folderId;
5438
5221
  if (params.tags) uploadBody.tags = params.tags;
@@ -5467,6 +5250,84 @@ ${JSON.stringify(uploadRes.data, null, 2)}` }],
5467
5250
  }
5468
5251
  }
5469
5252
  );
5253
+ registerSpeakTool(
5254
+ server,
5255
+ "upload_and_analyze_batch",
5256
+ `Upload several URLs in one call \u2014 the batch form of upload_and_analyze, for when someone hands you a list of links. Takes up to ${MAX_BATCH_URLS} URLs and starts at most ${MAX_BATCH_CONCURRENCY} at a time so a long list does not hammer the API. Each URL may be a direct/public file URL or a shareable social/video page link. Supported page links: ${SUPPORTED_URL_SOURCES}. ${UNSUPPORTED_URL_SOURCES} One URL failing does not stop the rest: every URL is reported individually as uploaded or failed, with its reason. Returns as soon as the uploads are accepted, so poll get_media_status per mediaId, or list_media on the folder, to follow processing. Prefer this over calling upload_and_analyze in a loop.`,
5257
+ {
5258
+ urls: import_zod15.z.array(import_zod15.z.string().min(1)).min(1).max(MAX_BATCH_URLS).describe("The URLs to import, up to 25. Pass each one exactly as the user gave it; page links are resolved server-side. Duplicates are uploaded once."),
5259
+ mediaType: import_zod15.z.enum([MediaType.AUDIO, MediaType.VIDEO]).optional().describe('Applies to every URL in the batch. Send it only when the user has said which they want for all of them \u2014 "audio" if they asked for audio only, "video" if they called them videos. Otherwise omit it and the server decides per URL. Mixed batches: leave it off, or split into two calls.'),
5260
+ folderId: import_zod15.z.string().optional().describe("Folder ID for every upload in the batch"),
5261
+ sourceLanguage: import_zod15.z.string().optional().describe('BCP-47 language code applied to every upload, e.g. "en-US"'),
5262
+ tags: import_zod15.z.string().optional().describe("Comma-separated tags applied to every upload"),
5263
+ concurrency: import_zod15.z.number().int().min(1).max(MAX_BATCH_CONCURRENCY).optional().describe("How many uploads to start at once, 1 to 5. Defaults to 5. Drop it to 1 to import strictly in order.")
5264
+ },
5265
+ {
5266
+ title: "Upload and Analyze Several URLs",
5267
+ readOnlyHint: false,
5268
+ destructiveHint: false,
5269
+ idempotentHint: false,
5270
+ openWorldHint: true
5271
+ },
5272
+ async (params) => {
5273
+ const urls = [...new Set(params.urls.map((u) => u.trim()).filter(Boolean))];
5274
+ if (urls.length === 0) {
5275
+ return { content: [{ type: "text", text: "Error: no usable URLs after trimming." }], isError: true };
5276
+ }
5277
+ const shared = {};
5278
+ if (params.mediaType) shared.mediaType = params.mediaType;
5279
+ if (params.sourceLanguage) shared.sourceLanguage = params.sourceLanguage;
5280
+ if (params.folderId) shared.folderId = params.folderId;
5281
+ if (params.tags) shared.tags = params.tags;
5282
+ const uploaded = [];
5283
+ const failed = [];
5284
+ let cursor = 0;
5285
+ const workerCount = Math.min(params.concurrency ?? MAX_BATCH_CONCURRENCY, urls.length);
5286
+ const send = (url) => api.post("/v1/media/upload", {
5287
+ ...shared,
5288
+ name: url.split("/").pop()?.split("?")[0] || "Upload",
5289
+ url
5290
+ });
5291
+ const worker = async () => {
5292
+ for (let i = cursor++; i < urls.length; i = cursor++) {
5293
+ const url = urls[i];
5294
+ try {
5295
+ let res;
5296
+ try {
5297
+ res = await send(url);
5298
+ } catch (err) {
5299
+ const message = formatAxiosError(err);
5300
+ if (!isRateLimited(message)) throw err;
5301
+ await new Promise((r) => setTimeout(r, RATE_LIMIT_RETRY_DELAY_MS));
5302
+ res = await send(url);
5303
+ }
5304
+ const mediaId = res.data?.data?.mediaId;
5305
+ if (mediaId) uploaded.push({ url, mediaId, state: res.data?.data?.state ?? "pending" });
5306
+ else failed.push({ url, error: "Upload accepted but no mediaId was returned." });
5307
+ } catch (err) {
5308
+ failed.push({ url, error: formatAxiosError(err) });
5309
+ }
5310
+ }
5311
+ };
5312
+ await Promise.all(Array.from({ length: workerCount }, worker));
5313
+ const result = {
5314
+ requested: urls.length,
5315
+ uploaded: uploaded.length,
5316
+ failed: failed.length,
5317
+ media: uploaded,
5318
+ errors: failed,
5319
+ nextSteps: uploaded.length ? [
5320
+ `1. Poll get_media_status for each of the ${uploaded.length} mediaId values every 10-30 seconds, or call list_media on the folder to see them together.`,
5321
+ `2. When one reads "processed", call get_media_insights and get_transcript for it.`,
5322
+ failed.length ? `3. ${failed.length} URL(s) did not upload \u2014 report the reasons above rather than silently retrying.` : `3. Nothing failed in this batch.`
5323
+ ] : ["No uploads were accepted. Report the errors above to the user."]
5324
+ };
5325
+ return {
5326
+ content: [{ type: "text", text: JSON.stringify(result, null, 2) }],
5327
+ ...uploaded.length === 0 ? { isError: true } : {}
5328
+ };
5329
+ }
5330
+ );
5470
5331
  registerSpeakTool(
5471
5332
  server,
5472
5333
  "upload_local_file",
@@ -5501,11 +5362,10 @@ ${JSON.stringify(uploadRes.data, null, 2)}` }],
5501
5362
  };
5502
5363
  }
5503
5364
  const filename = path2.basename(filePath);
5504
- const isVideo = isVideoFile(filePath);
5505
5365
  const mediaType = params.mediaType ?? detectMediaType(filePath);
5506
5366
  const mimeType = getMimeType(filePath);
5507
5367
  const signedRes = await api.get("/v1/media/upload/signedurl", {
5508
- params: { isVideo, filename, mimeType }
5368
+ params: { mediaType, filename, mimeType }
5509
5369
  });
5510
5370
  const signedData = signedRes.data?.data;
5511
5371
  const uploadUrl = signedData?.preSignedUrl ?? signedData?.signedUrl ?? signedData?.url;
@@ -5561,7 +5421,7 @@ ${JSON.stringify(signedRes.data, null, 2)}` }],
5561
5421
  }
5562
5422
  );
5563
5423
  }
5564
- var import_zod15, fs, path2, CANONICAL_FILTER_FIELDS, ID_PATTERN, TRIGGER_SPEC_DESCRIPTION, STEP_SPEC_DESCRIPTION, buildAutomationSchema;
5424
+ var import_zod15, fs, path2, MAX_BATCH_URLS, MAX_BATCH_CONCURRENCY, RATE_LIMIT_RETRY_DELAY_MS, isRateLimited, CANONICAL_FILTER_FIELDS, ID_PATTERN, TRIGGER_SPEC_DESCRIPTION, STEP_SPEC_DESCRIPTION, buildAutomationSchema;
5565
5425
  var init_workflows = __esm({
5566
5426
  "src/tools/workflows.ts"() {
5567
5427
  "use strict";
@@ -5574,6 +5434,10 @@ var init_workflows = __esm({
5574
5434
  init_media_utils();
5575
5435
  init_capabilities();
5576
5436
  init_inbound_webhook_utils();
5437
+ MAX_BATCH_URLS = 25;
5438
+ MAX_BATCH_CONCURRENCY = 5;
5439
+ RATE_LIMIT_RETRY_DELAY_MS = 5e3;
5440
+ isRateLimited = (message) => /rate limit|too many requests|\b429\b/i.test(message);
5577
5441
  CANONICAL_FILTER_FIELDS = /* @__PURE__ */ new Set([
5578
5442
  "name",
5579
5443
  "duration",
@@ -6408,7 +6272,7 @@ function register16(server, client) {
6408
6272
  readOnlyHint: false,
6409
6273
  destructiveHint: false,
6410
6274
  idempotentHint: true,
6411
- openWorldHint: true
6275
+ openWorldHint: false
6412
6276
  },
6413
6277
  async ({ dashboardId }) => {
6414
6278
  try {
@@ -6696,7 +6560,7 @@ function registerPrompts(server) {
6696
6560
  "analyze-meeting",
6697
6561
  "Upload a meeting recording and get a full analysis \u2014 transcript, insights, action items, and key takeaways.",
6698
6562
  {
6699
- url: import_zod18.z.string().describe("URL of the meeting recording \u2014 a direct file link or a shareable social/video link (resolved automatically)"),
6563
+ url: import_zod18.z.string().describe(`URL of the meeting recording \u2014 a direct file link, or a shareable page link from ${SUPPORTED_URL_SOURCES} (resolved to the underlying media automatically)`),
6700
6564
  name: import_zod18.z.string().optional().describe("Meeting name (optional)")
6701
6565
  },
6702
6566
  async ({ url, name }) => ({
@@ -6718,7 +6582,10 @@ function registerPrompts(server) {
6718
6582
  ` - Open questions or follow-ups needed`,
6719
6583
  ` - Overall sentiment`,
6720
6584
  ``,
6721
- `Use upload_and_analyze to handle the upload and processing in one step.`
6585
+ `Use upload_and_analyze to handle the upload and processing in one step. Pass the URL`,
6586
+ `exactly as given \u2014 a page link is resolved server-side. If I told you whether this is`,
6587
+ `an audio or a video recording, pass that as mediaType; if I did not say, leave it off`,
6588
+ `and let the server pick the best available track.`
6722
6589
  ].join("\n")
6723
6590
  }
6724
6591
  }
@@ -6806,6 +6673,7 @@ var init_prompts = __esm({
6806
6673
  "src/prompts.ts"() {
6807
6674
  "use strict";
6808
6675
  import_zod18 = require("zod");
6676
+ init_media_utils();
6809
6677
  }
6810
6678
  });
6811
6679
 
@@ -6931,6 +6799,7 @@ var init_tool_names = __esm({
6931
6799
  // workflows (high-level wrappers around media + upload + automation tools)
6932
6800
  "build_automation",
6933
6801
  "upload_and_analyze",
6802
+ "upload_and_analyze_batch",
6934
6803
  "upload_local_file",
6935
6804
  // users / team management
6936
6805
  "list_users",
@@ -7387,11 +7256,10 @@ function createCli() {
7387
7256
  let state;
7388
7257
  if (isLocalFile) {
7389
7258
  const filename = pathMod.basename(source);
7390
- const isVideo = isVideoFile(source);
7391
7259
  const mediaType = opts.type ?? detectMediaType(source);
7392
7260
  const mimeType = getMimeType(source);
7393
7261
  const signedRes = await client.get("/v1/media/upload/signedurl", {
7394
- params: { isVideo, filename, mimeType }
7262
+ params: { mediaType, filename, mimeType }
7395
7263
  });
7396
7264
  const signedData = signedRes.data?.data;
7397
7265
  const uploadUrl = signedData?.signedUrl ?? signedData?.url;
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@speakai/mcp-server",
3
- "version": "1.23.0",
3
+ "version": "1.24.1",
4
4
  "mcpName": "io.github.speakai/mcp-server",
5
5
  "description": "Official Speak AI MCP Server — capture meetings, search thousands of recordings, run async voice and video surveys, create clips, and automate workflows from your AI assistant.",
6
6
  "homepage": "https://docs.speakai.co/mcp",