Transcription in n8n: Built-in Nodes vs Community Nodes

n8n can transcribe audio without any community node: the built-in OpenAI node has a Transcribe a Recording operation, and the AWS Transcribe node runs transcription jobs on Amazon. Community nodes add other speech-to-text services, transcription on your own server, and nodes that fetch transcripts of YouTube and other videos instead of transcribing audio. This guide compares them, using n8n’s documentation, the description and README each author publishes on npm and our own download data.

Built-in transcription

• OpenAI, Transcribe a Recording: per n8n’s documentation, it takes an audio file from a binary field (.flac, .mp3, .mp4, .mpeg, .mpga, .m4a, .ogg, .wav or .webm), uses the whisper-1 model by default, and is limited to 25 MB per file by the OpenAI API. A Translate a Recording operation turns speech into English text.
• AWS Transcribe: create, get, get all and delete transcription jobs.

Transcription community nodes

Our Transcription hub lists 28 packages that pass our quality rule. The most downloaded, with the description each author gives on npm (weekly npm downloads as of 28 October 2026):

Versions and dates come from the npm registry. Weekly downloads: npm, last 7 days, adjusted for npm outage days.

n8n-nodes-transcript-lol

Published by Transcript.lol itself, and by far the most downloaded package in this group. The action node creates recordings for transcription, gets one or many recordings, lists your workspaces and gets a transcript as JSON, plain text, CSV, Word, SRT or VTT. A trigger node starts a workflow when a transcript is completed or fails, when new AI insights are generated, or when a translation is completed. Sign-in uses OAuth2 with PKCE.

Other speech-to-text services

• n8n-nodes-nvidia-nim-whisper-v2: Whisper Large V3 on NVIDIA NIM, to transcribe or translate to English. Its README lists WAV (mono 16-bit), OPUS and FLAC, automatic language detection or 30+ language codes, word time offsets, and both NVIDIA’s cloud API and self-hosted NIM, with an NVIDIA API key.
• n8n-nodes-openrouter-transcribe: sends audio (binary or base64) to an OpenRouter model that accepts audio input, openai/gpt-4o-audio-preview by default, with an optional prompt. Formats include WAV, MP3, AIFF, AAC, OGG, FLAC and M4A.
• n8n-nodes-vatis-transcribe: from Vatis Tech’s GitHub organization; uploads media from a binary field or a URL, polls the processing status and downloads the transcript as structured JSON.

n8n-nodes-transcribe-audio (on your own server)

Runs Whisper on your n8n server, so no transcription API key is needed and, per its README, the audio isn’t sent to a transcription service. Its README sets clear limits: self-hosted n8n only, not n8n Cloud; WAV and MP3 files with English speech; the first run downloads the speech model, so the server needs internet access; and transcription runs on the server’s CPU, so longer recordings and larger models take more time and memory.

Video transcripts instead of audio

• n8n-nodes-t3data-youtube-transcripts: transcripts of YouTube videos, Shorts, playlists or whole channels as text, timestamped segments, SRT or chunks for RAG. It runs an Apify Actor with your Apify token; its README gives the price as 5 USD per 1,000 transcripts, with videos without captions free. It also works as an AI Agent tool.
• n8n-nodes-transcriptfetch: transcripts from YouTube, TikTok and Instagram through the TranscriptFetch API, with AI speech-to-text when a video has no captions, plus YouTube search, channel and playlist listing and a new-video trigger.

Check each node page here for the latest release and, where it has one, its GitHub repository for open issues before you build on a package.

How to choose

• You already use OpenAI and your files are under 25 MB: the built-in OpenAI node, with nothing to install.
• Audio that must stay on your server: n8n-nodes-transcribe-audio, for English WAV or MP3.
• Whisper Large V3 or a self-hosted NIM: n8n-nodes-nvidia-nim-whisper-v2.
• A Transcript.lol account, or workflows that start when a transcript is ready: n8n-nodes-transcript-lol.
• The transcript of a YouTube video, playlist or channel: n8n-nodes-t3data-youtube-transcripts or n8n-nodes-transcriptfetch, without downloading the audio.

To install a package on a self-hosted instance, an owner or admin enters its npm name in Settings > Community Nodes; full steps are in How to Install n8n Community Nodes. On n8n Cloud only verified community nodes can be installed, see Verified vs Unverified n8n Community Nodes. Related: YouTube and n8n and FFmpeg in n8n.

FAQ

Can n8n transcribe audio?

Yes. The built-in OpenAI node’s Transcribe a Recording operation turns an audio file into text, and the AWS Transcribe node runs jobs on Amazon. Community nodes add other services and local transcription.

Can I transcribe audio in n8n without an API key?

On a self-hosted instance, n8n-nodes-transcribe-audio runs Whisper on your own server. Its README limits it to English speech in WAV and MP3 files.

How do I get a YouTube transcript in n8n?

Use a node that fetches existing transcripts, such as n8n-nodes-t3data-youtube-transcripts (through Apify) or n8n-nodes-transcriptfetch, instead of transcribing the audio yourself.

What is the most used transcription community node?

n8n-nodes-transcript-lol, with 13,545 weekly npm downloads on 28 October 2026.