Comparison

Nojoin Next to Jamie, Otter and Granola

Nojoin has taken no funding, employs nobody, and sells nothing you can't download. Here it is next to three products that raised money to build the same thing.

Every claim below comes from that vendor's own current documentation. Where they're better, the table says so — which is the part most comparison pages leave out.

At a glance

	Nojoin	Jamie	Otter	Granola
Runs on your hardware, and stays there	Yes. One compose file	No. Jamie's EU cloud	No. Otter's US cloud	No. Granola's US cloud
Open source, and your choice of model	Yes. AGPLv3, your keys or Ollama	No. Proprietary, vendor-chosen	No. Proprietary, vendor-chosen	No. Proprietary, vendor-chosen
No bot joins the call	Yes. Any browser, never a participant	Yes. Captures locally	Partly. Bot by default	Yes. Captures locally
Speakers remembered between meetings	Yes. Voiceprint library	Yes. Named once, then reused	Yes. Auto-tags names	No. Per-meeting tags
Voice models you hold and can rebuild	Yes. Recalibrate, rebuild, delete	No. Not stated in docs	No. Not stated in docs	No. Not stated in docs
An assistant can change the record	Yes. Thirty tools, your server	Partly. Creates and applies tags	No. Reads Otter, writes elsewhere	No. Reads notes and transcripts
Your data comes out whole	Yes. Documents, audio, backups	Partly. Syncs to apps, no download	Yes. Documents and audio	Partly. CSV of notes
Nothing runs out	Yes. No plans, no meters	Partly. Credits, and a length cap	Partly. Capped below Business	Partly. History capped on free

Detailed comparison

	Nojoin	Jamie	Otter	Granola
How audio is captured	Your browser captures the call. Nothing joins the meeting or appears in the participant list.	A native macOS or Windows app captures audio on your machine, online or in the room. No bot joins the call.	An AI agent joins Zoom, Teams and Meet calls. Bot-free recording is also available from the desktop, mobile or web app.	No bot. Desktop apps for macOS and Windows, plus an iPhone app, capture system audio and microphone.
Speaker attribution	A speaker library you can open. Voiceprints recalibrate from better samples and rebuild after an upgrade.	Name each speaker once from an audio clip; Jamie identifies them in later meetings. No voiceprint library documented.	Speaker identification recognises tagged speakers and auto-tags them in later transcripts.	Speaker tags come from platform display names on Meet and Zoom. No cross-meeting memory documented.
Where processing happens	On your own server.	Entirely in the EU: audio to Frankfurt, transcribed on Modal's GPUs, notes through the Anthropic or OpenAI APIs.	In Otter's cloud, principally on AWS in the United States.	Audio goes from your device to Granola's cloud transcription and AI providers.
Where recordings live	On your own server.	Transcripts sit on a server in Frankfurt; audio is deleted permanently once the transcript is ready.	In Otter's cloud on AWS S3, with server-side encryption.	Audio is deleted after transcription; transcripts and notes are stored on US-hosted AWS.
Self-hosting	Self-hosting is the product: one compose file on your hardware.	None documented; the app is local but every meeting is processed and stored in Jamie's EU cloud.	Not stated in current docs	None documented; the service runs in Granola's US-hosted cloud.
Source and licence	Open source under AGPLv3.	Closed source: a non-transferable right to use, lasting as long as the contract.	Closed source, proprietary licence.	Closed source, proprietary licence.
Model choice	Bring your own API keys, or run fully local inference with Ollama.	Vendor-chosen: notes come from the Anthropic or OpenAI APIs. No bring-your-own-key or local option documented.	Otter's own AI plus vendor-chosen third-party providers; no bring-your-own-key or local option documented.	Vendor-chosen providers (Deepgram, AssemblyAI, OpenAI, Anthropic); no bring-your-own-key or local option documented.
Live in-meeting guidance	Meeting Edge: live questions worth asking, missed points, and concept help.	The in-meeting widget is a scratchpad for your own notes. Notes and transcript come after you stop recording.	The meeting agent gives real-time transcription, takeaways and action items, and answers questions live.	Granola Chat can answer questions about the ongoing meeting.
Calendar integration	Google and Outlook sync, with meeting context in the dashboard.	Google Calendar or Outlook, for reminders, meeting titles and better speaker identification.	Google Calendar, Microsoft Outlook and iOS Calendar.	Google and Microsoft/Outlook; Apple Calendar isn't supported.
Search	One query across recordings, notes and documents.	Semantic search across your meeting content, from the app, the API and the MCP server.	Search and AI chat across meetings and connected apps, by keyword, speaker and date.	Chat spans notes and transcripts across meetings and folders, plus files you upload.
What an assistant can do with your meetings	Thirty tools on your deployment: correct a line, name a speaker, re-run the notes, file the tasks. Every change is labelled and reversible.	Thirteen tools for Claude, ChatGPT, Cursor and Copilot. Nine read meetings, transcripts, tasks and people; four create and apply tags.	Reads transcripts to search, analyse themes and generate content. Writes outward into Docs, Jira, Salesforce and Slack. Otter for Enterprise.	Reads notes and transcripts on paid plans: list, read and search meetings. On Enterprise it's beta, off until an admin enables it.
Limits on how much you process	None. No allowance, no per-meeting ceiling, no history that expires. What you get through is a hardware question.	One credit a meeting. 10 a month and 30-minute meetings free; 20 and two hours next. Higher plans drop the count, keep a three-hour cap.	300 minutes a month, 30 per conversation, on the entry plan; 1,200 and 90 next. Upper plans lift the monthly cap but keep a four-hour ceiling.	Meeting history is limited on the entry plan, by an amount the pricing page doesn't state. Unlimited above it.
Export and portability	Transcript, notes or both as TXT, PDF or DOCX, timestamped per line, plus MP3 audio and full-instance backups.	Summary, transcript and tasks sync into Notion, Google Docs, HubSpot, Salesforce, Asana and more, with an API and webhooks. No file download documented.	TXT, DOCX, PDF and SRT, plus audio export.	Bulk export of notes to CSV: titles, summaries, transcripts and details.