Otter AI Alternative No Bot: The Audio Path Decides
By Kyle Nelson, Founder, Fazit
ogTitle: "Otter AI Alternative No Bot: Where Does the Audio Go?"
Otter.ai now offers bot-free recording through its Mac or Windows desktop app, and it says the same is available through Chrome. That answers the "no bot" half of your search. It does not answer the question that matters for a consultant, lawyer, financial adviser or therapist on a confidential 1:1 call: where does the audio go after you click start?
Fazit is a native macOS menubar app that turns 1:1 client calls into finished Markdown notes in an Obsidian vault. No bot joins the call, transcription and note generation run on the Mac, and the call audio is never written to disk. The difference is architectural, and it survives a security review in a way that a bot-free label does not.
Otter AI alternative no bot
Otter's own site confirms the bot-free option: you can record conversations directly from a desktop, mobile device or browser, with live transcription and speaker recognition. The AI Meeting Agent can also join and record meetings on Zoom, Microsoft Teams and Google Meet. So Otter gives you two capture modes: a visible bot that joins, or a desktop app that records without one.
The second mode is what people searching for "otter ai alternative no bot" usually want. They want the note without the third participant in the waiting room. Otter delivers that. What it does not deliver is a call where the audio stays on your machine.
Otter's free plan includes 300 monthly transcription minutes, with each conversation limited to 30 minutes and three audio or video file imports. The paid plan advertises 6,000 minutes and unlimited file imports. Those numbers tell you the product's center of gravity: audio files, stored and processed in Otter's cloud.
Bot-free is not the same as private
The 2026 botless notetaker wave made this confusion common. Granola, Krisp, Circleback, and now Fireflies and Fathom all capture meetings without a bot. But botless describes what your meeting looks like, not where your audio goes. The audio still streams to vendor clouds and third-party speech services. They removed the bot. They did not remove the server.
I wrote about this in Bot-Free Is Not the Same as Private: The 2026 Botless Notetaker Wave. The distinction matters because a client sees the participant list, not the network traffic. A call can look clean and still create a recording on someone else's infrastructure.
Otter's bot-free mode is the same shape. No bot appears in the call, but Otter's documentation describes recording, transcription and speaker recognition running through its platform. The audio reaches Otter's systems. For a therapist with a session note obligation, or a lawyer with privilege concerns, that is the fact that survives the demo.
What Otter's wiretap lawsuit changed
Otter.ai faces a consolidated federal class action, In re Otter.AI Privacy Litigation, in the Northern District of California. The complaint alleges that OtterPilot records meetings without everyone's consent and feeds the audio into model training. No court has ruled yet. A complaint is not a finding.
But the case gave the category a price. Under California Penal Code § 637.2, the claim is $5,000 per recording. I covered the case file in Brewer v. Otter.ai: The Case File, and Where It Stands. The August 4 hearing did not go Otter's way on standing, and the case is now the one the whole category watches.
The legal risk points at one artifact: the recording. A transcript that exists because a recording existed becomes discoverable evidence in any later dispute. The one you never made cannot be subpoenaed. That is not a marketing line. It is the difference between a pipeline that creates a durable audio file and one that never does.
Where Fazit draws the line
Fazit captures two audio streams on the Mac: your microphone and the call app's own output. A Core Audio process tap reads the call app directly. Nothing joins the meeting and no other participant sees anything. The audio exists only in a fixed-size RAM ring buffer with no write API, destroyed on every exit path including errors.
Transcription runs on-device using Parakeet via CoreML. The note is generated by a local model on localhost. The output is a plain Markdown file written into a folder of your Obsidian vault, or Apple Notes. There is no Fazit database holding a copy.
Every note records audio_retained: false in its frontmatter. That is a statement about the pipeline, not a promise about your conduct. Consent and professional-conduct duties still apply to the person running the call. This is not legal advice.
The comparison that matters
| Tool | Bot joins? | Audio leaves the Mac? | Audio file created? | Note location |
|---|---|---|---|---|
| Otter.ai (bot mode) | Yes | Yes | Yes | Otter cloud |
| Otter.ai (bot-free mode) | No | Yes | Yes | Otter cloud |
| Granola | No | Yes | Yes, deleted after transcription | Granola cloud |
| Fazit | No | No | No | Your Obsidian vault |
Otter's bot-free mode removes the visible third party. It does not remove the server. Granola's public security FAQ names OpenAI and Anthropic as subprocessors, uses US-only AWS storage, and retains transcripts indefinitely by default. I detailed that in Granola Subprocessors: What the Public Docs Actually Say.
Fazit is the row where the audio file was never eligible to exist. That is the column that survives a security review, and it is the one a client's NDA or a regulator's question will point at.
What you actually get in the note
When a Fazit capture ends, a note appears in <vault>/Fazit/. It contains a summary, key moments, action items with owners, and a paste-ready follow-up email in the client's own language and register. The transcript is attributed as You: and Them: because the two audio streams are physically separate. On a 1:1 call, that attribution is exact without diarization guesswork.
The note is a file you can grep, move, back up and delete. It is not a row in a vendor's database. I wrote about why that matters in Obsidian Meeting Notes Should Be Files, Not Rows in a Vendor's Database.
For a consultant whose MSA names AI tools outright, the note workflow is the easy half. The audio path is the half that decides whether the tool satisfies the confidentiality clause. A cloud notetaker on a client call is a disclosure to a third party. A local pipeline that never creates an audio file is not.
The consent question does not disappear
Fazit's architecture reduces the vendor exposure and the audio-file exposure. It does not remove your duty to have consent where the law requires it. California's CIPA is all-party. Germany's § 201 StGB targets recording non-public speech, and German counsel's view is that a RAM-only pipeline that never writes an audio file generally does not meet the elements of the offense because there is no recording on a sound carrier. That is a legal analysis, not a guarantee, and it is not legal advice.
Before a call, say the honest line: the audio is not kept, and the other person can decline. That gives the other participant a real choice, which is what consent law is for.
Otter.ai's bot-free mode is a real feature. It solves the visible-bot problem. It does not solve the recording problem, because the recording still exists on Otter's infrastructure. For confidential 1:1 client calls, the alternative that answers the search query fully is the one where no bot joins and no audio file is ever created.