# TranscribeNext (https://transcribenext.com) > AI transcription platform that turns audio and video into editable, searchable > text across 100 languages, with automatic speaker detection and timestamps. > Web app plus a native Mac app. Founded 2023. Every figure below is taken from the product's own configuration. Claims that could not be verified against it are deliberately absent. ## Key pages - https://transcribenext.com: main platform - https://transcribenext.com/pricing: plans and limits - https://transcribenext.com/features: capabilities - https://transcribenext.com/stats: full supported-language list with accuracy tiers - https://transcribenext.com/formats: per-format transcription guides - https://transcribenext.com/use-cases: applications by audience - https://transcribenext.com/blog: guides and comparisons - https://transcribenext.com/faq: frequently asked questions - https://transcribenext.com/help: help centre ## Tools Pages where a file can be uploaded and transcribed directly, each with a guide: - https://transcribenext.com/transcribe-audio-to-text: audio to text, step by step - https://transcribenext.com/mp4-to-transcript: video (MP4, MOV) to transcript - https://transcribenext.com/convert-audio-to-text: audio file converter - https://transcribenext.com/transcription-with-timestamps: timestamp formats (SRT, VTT, TXT, DOCX) and measured timestamp accuracy - https://transcribenext.com/german-speech-to-text: German transcription, with a published engine comparison - https://transcribenext.com/mac-speech-to-text: speech to text and dictation on a Mac - https://transcribenext.com/download: the native Mac app - https://transcribenext.com/mcp-server: connect ChatGPT, Claude, Cursor or VS Code to your transcripts (MCP server) ## Published data Original measurements, with the method and sample size stated on the page: - Transcription timestamp accuracy on AMI meeting audio (69 minutes, 1,373 annotated utterance starts, CC BY 4.0): https://transcribenext.com/data/timestamp-accuracy-ami-2026-09.csv https://transcribenext.com/data/timestamp-accuracy-ami-2026-09.json described at https://transcribenext.com/transcription-with-timestamps#measured ## Connect an AI assistant (MCP) MCP server address: https://mcp.transcribenext.com/mcp (HTTP transport, OAuth sign-in; Cursor and other clients without sign-in use an API key from Integrations → AI & API). Works with ChatGPT, Claude (claude.ai, Claude Desktop, Claude Code), Gemini, Grok, VS Code, Cursor, Codex, Zed and goose. A free account is enough. Tools: search_transcripts, list_transcripts, get_transcript, get_summary and list_folders_and_tags read; transcribe_url (Google Drive, Dropbox, YouTube links), upload_file and finish_upload (a file from the user's computer, in apps that run commands) start transcriptions. No tool edits or deletes anything. New transcriptions started through AI assistants: Free 3 a day and Plus 20 a day (shared with the website), Premium 50 a day (no daily limit on the website); reset at 00:00 UTC. More by arrangement: info@transcribenext.com, billed by the hour. Page: https://transcribenext.com/mcp-server Setup guide: https://transcribenext.com/help/integrations/connect-chatgpt-claude ## What it does Upload an audio or video file, paste a YouTube link, or record a meeting from Zoom, Google Meet or Microsoft Teams. The language is detected automatically, speakers are labelled, timestamps are added, and the transcript is ready to edit, search and export. ## Language coverage 100 languages, published in four accuracy tiers rather than as a single figure: - 15 languages in the highest-accuracy tier (English, Chinese, Spanish and other heavily-resourced languages) - 25 European languages - 30 Asian and other languages - 30 specialised and lower-resource languages Accuracy depends on which tier the audio falls in, and on recording quality. ## Input formats Accepted directly, with no conversion needed: - Audio: MP3, WAV, M4A, AAC, FLAC, OGG, WMA - Video: MP4, MOV, WEBM, MKV, MPEG, AVI Video is accepted as-is; the audio track does not need extracting first. ## Speaker detection Speaker count can be left on automatic or pinned to an exact number from 2 to 20. Pinning is more reliable when the number of participants is known. ## Plans Free — $0 - 3 files per day, 5-minute transcript preview per file (preview only) - 99% accuracy, max 50 MB per file, 150 MB storage - Export to TXT, Markdown, JSON - No credit card required Plus — $19.99/month, or $8.33/month billed annually ($99.99/year) - 20 files per day, up to 5 hours per file, full transcription - 99.5% accuracy, speaker detection, AI summaries and action items - Meeting recording (Zoom, Google Meet, Microsoft Teams) - Max 2 GB per file, 100 GB storage - Export to TXT, Markdown, JSON, PDF, DOCX, SRT, VTT, or the original audio Premium — $29.99/month, or $12.49/month billed annually ($149.99/year) - Unlimited files per day, up to 8 hours per file - Everything in Plus, plus calendar integration and priority support - Max 5 GB per file, 1 TB storage Billing is available in USD, EUR, GBP and JPY. ## Export formats TXT, Markdown, JSON on every plan; PDF, DOCX, SRT, VTT and the original audio on paid plans. SRT and VTT are subtitle formats accepted by YouTube's caption uploader. ## Identity - Wikidata: https://www.wikidata.org/wiki/Q140388965 - LinkedIn: https://www.linkedin.com/company/transcribenext - G2: https://www.g2.com/products/transcribenext/reviews - Capterra: https://www.capterra.com/p/10048868/TranscribeNext-com/