Meeting Transcription

Meeting mode detects 2-10 speakers automatically, generates an AI summary + bullet-point action items at the top, and extracts follow-ups + decisions + attendees so team leads can paste the output straight into their weekly update.

Bincike-da-kanka daga 2-10. Idan kana da sani game da adadin masu magana, ka daidaita slider don samun bayani mafi daidai — shawarwari masu bayyanawa suna rage kuskure na diarization da ~15% akan sauti mai zafi na ɗakin taron.

A 3-5 maganar shugabanci summary, a bullet jerin abubuwa na aiki da assignees lokacin da aka ambaci, muhimmanci shawarwari da aka yi, da sunaye masu halarta (daga mai magana da labels). Created by an LLM a kan saman na raw transcript - ~ 300 karin tokens.

Otter caps a minti 300 kyauta, $ 16.99 / watanni don minti 1,200, Ingilishi-mai nauyi. Muna yin harsuna 99 a ~ $ 0.003 / min biyan kuɗi-da-aiki, da kuma LLM-da aka samar da bayani + ayyukan abubuwa an haɗa su cikin farashin ~ 300-token AI bayan-process.

Fireflies shine $ 10 / mai amfani / watan ($ 18 / mai amfani don nazarin AI). Muna biyan kuɗin minti na minti, ba lasisin wurin zama ba, kuma mai fitar da nazarinmu yana kira da irin wannan darasin LLMs (Qwen 7B default, Claude / GPT upgradeable).

Gong is sales-call-focused at $1,200+/seat/year with CRM integrations. We do the transcription + summary + action items for any meeting type at a fraction of the cost — no CRM, but you get clean exports to paste anywhere.

Export as DOCX (Word-compatible, imports to Google Docs), TXT with speaker-labeled blocks, or JSON. Copy the summary + action items and paste into Notion — they render as a markdown list.

An goge fayiloli da aka tura bayan an yi waƙa. Waƙoƙin da aka saukar suna zaune a cikin asusunka na kwanaki 7 (da aka biya) ko sa'o'i 24 (ba da kuɗi). Ga tattaunawa masu sirri, ka goge layin tarihi bayan an saukar da su.

Na'am, amma daidaitawar da'ira ita ce mafi kyau idan kowanne mai halarta daga wuri mai nisa yana da miƙaƙƙofin sa. Sauti na ɗakin miƙaƙƙofin kawai da mutane da yawa suna magana na iya haɗa sau da yawa murya biyu cikin wata alamun shafi guda.

Mai kyau don alkawurran bayyanannu ("Sarah, aika da jirgin sama ta ranar Jumma'a") - 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90% na 80-90%

Na'am, amma /transcribe/zoom/ an gina shi ne don wani dalili — yana karanta sunaye na masu magana da aka rufe cikin metadata na Zoom ta atomatik don haka labels su zama "Alice / Bob" a matsayin "Mai magana 1 / 2".

Har zuwa 2GB (kimanin 12+ sa'o'i na 128 kbps sauti). Don rarrabawa mafi tsawo, raba zuwa sassa ko amfani da /api/ /batch.