Best Free Transcription Software: 6 Tools Tested and Reviewed [2026]

I compared 6 the best free transcription software; Jamie, Happy Scribe, oTranscribe, Otter.ai, TurboScribe and OpenAI Whisper.

The comparison was based on six areas: data security, transcription, speaker identification, free plan limits, supported languages, and price.

We also ran our own accuracy test on Jamie against 4 other AI note takers. Jamie achieved 98 out of 100 for identifying speakers and 90 out of 100 for transcribing audio; it ranked first of the five.

There are many options when you search for a free transcription tool. The catch is that the monthly allowance a vendor advertises sits on top of a per-session cap and what they offer in each plan.

And almost none of them actually tell you what their "pretty" pricing plan actually costs you in the long run, when even the tiniest feature keeps costing you credits.

Or maybe you don't even know how secure these AI note takers are, where your audio might end up going; there are many questions to be asked and answered.

But one of the questions remains the same, through and through: what is the best free transcription software? In this article, we are going to show you 6 of them.

‍

TL;DR

• Jamie: the best overall free transcription software, with consent first, bot-free AI notes and transcriptions for over 100 languages on Mac, Windows, iOS and Android. Jamie identifies each speaker by name when you start the recording, and Speaker Memory matches that same voice in later meetings. Ask AI answers a question across all your meetings. All meeting data is stored in Europe, and in addition to being GDPR compliant, Jamie is ISO 27001 certified and DORA-ready. Meeting audio is deleted when transcription is completed. All of these options are available in the free version, which allows you to have 10 meetings per month, at up to 30 minutes each. If you need more than that, then it is 25 euro per month for Plus with Agentic Intelligence and Plug and Play Integrations.
• Happy Scribe is the free transcription software for human proofreading, designed to produce AI generated transcripts in over 150 languages. After the AI produces the first draft of the transcript, it may be reviewed by a human editor. Additionally, Happy Scribe's platform provides several features which support the reviewer, including automatic speaker detection, which identifies each speaker for you, Action Notes, which highlight follow ups, tasks and deadlines after the meeting, a bot free browser recorder, and a Notetaker which joins the meeting as a participant instead. In addition, Happy Scribe has achieved SOC 2 Type 2 certification, and their security page says their data centre is located in the European Union. The free plan allows unlimited meetings at 45 minutes each, and Basic is 17 dollars a month.
• oTranscribe is the best free transcription software option for typing an interview yourself, with a media player on a web page, timestamps and keyboard shortcuts, in over twenty different interface language options. Your interviews will be exported from oTranscribe in either plain text, Markdown or .otr format, and your transcript can also be sent to Google Drive. The oTranscribe application can be used regardless of whether you have internet access. Your audio and transcript never leave your computer, no login information is needed to start using oTranscribe, and oTranscribe does not cost anything however often you use it.
• Best free transcription software for live meeting notes: Otter.ai, a Notetaker that joins your meetings via Zoom, Microsoft Teams or Google Meet, plus a desktop application to record the same meetings without any bot. It uses speaker identification that learns a voice you tag, provides AI Chat across meetings, and has SOC 2 Type 2 behind it. There are some limits on the free version, 300 minutes of transcription per month, as well as on each recording, which is capped at 30 minutes, while Pro is 16.99 dollars a month.
• TurboScribe is the best free transcription software to transcribe your uploaded audio or video files, with speaker recognition and 98+ languages of transcription on files that are already yours. You can also upload using a web link or record directly through your computer microphone. Translated output is available in 134+ languages. TurboScribe will export your completed transcripts as DOCX, PDF, TXT, SRT or VTT, and batch export in every format is a TurboScribe Unlimited feature. No training data will be made from the files you upload. The free option will allow you to create 3 transcripts per day at up to 30 minutes each. If you want more than that, you will need Unlimited, which costs 20 dollars a month.
• OpenAI Whisper is the best free transcription software to provide unlimited local transcription. An MIT licensed open source model from OpenAI that includes both the code and the weights to run 99 languages on your own computer, without uploading anything and without a monthly limit. The same command will also translate your speech into English. Whisper is free no matter how much you use it. The cost is learning a command line and giving up a few gigabytes of memory.

‍

What Are The Best Free Transcription Software?

The best free transcription software are Jamie, Happy Scribe, oTranscribe, Otter.ai, TurboScribe, and OpenAI Whisper. Each one will convert the speech in an interview or a meeting into a written document that you can read, search, and export. Each of the above options allows you to use their service at no cost.

Jamie generates the Executive level AI notes, an accurate, speaker-assigned, and labelled transcript, the action items, and the speaker names from a recording made directly on your own device in under 30 seconds, on every plan, including free. All meeting data generated by Jamie is stored in Europe, and Jamie is GDPR compliant as well as ISO 27001 certified.

Happy Scribe is also located within a European data centre, and in addition to creating the AI generated transcript, they add the element of having a real person proofread it. oTranscribe and OpenAI Whisper store the audio file locally on your computer. Otter.ai identifies the speakers in each of your meetings and allows you to ask questions about several meetings at once, just like Jamie's Ask AI's functionality. TurboScribe can handle over 98 different languages when you upload files.

Below is the comparison chart of the six free transcription software tools that I compared after thorough research. Clicking a cell will show the full answer for Jamie, Happy Scribe, oTranscribe, Otter.ai, TurboScribe and OpenAI Whisper.

What Each Free Plan Actually Gives You, Per Session And Per Month

Every one of these free transcription software adverts says "Unlimited" , on their free plan, and I want to see how truly unlimited and generous that unlimited free plan is and how other features could be locked behind the unlimited free plan. Especially considering how those free plans could be restricting one's productivity in the workplace. This applies to all free plans of the transcription softwares in the article, including ours. My goal here is to find the most generous free plan that you can try using today.

1. Jamie: Best Free Transcription Software for Bot Free Meeting Notes

Best for: Jamie is the free transcription software for teams who want the meeting audio deleted after transcription, GDPR and ISO 27001 in our own name, 100+ languages, Ask AI, and every feature on the free plan.

Similar to: Otter.ai, Happy Scribe, Fireflies.ai and MeetGeek.

Jamie is an AI powered bot free digital note taker, as well as free transcription software. Jamie transcribes and takes notes on meetings from your own device. Jamie generates the notes, transcript and action items automatically on the device used to make the recording.

We created Jamie this way so that the recording would never appear as a guest on someone else's conference call. Meeting data collected using Jamie is always stored and processed within Europe, and Jamie is both GDPR compliant and ISO 27001 certified.

Jamie Generates The Summary, The Full Transcript And The Action Items After Every Meeting

From each meeting recorded using Jamie, Jamie generates a summary organised by topics, a full transcript and a list of action items. Within seconds after the conclusion of each meeting, Jamie generates Executive Level AI Notes in all three parts: the summary, the transcript and the action items, and we place those on every plan including the free plan.

• Executive Level AI Notes: The summary is broken down into an executive summary and a full summary, and Jamie generates both in under 30 seconds. The summaries are formatted by topic, quote, action item, due date, priority level and speakers identified, with a template applied automatically based on the type of meeting.
• Live Transcript: Jamie offers live transcription. If you want to turn this option on, you go to the Sidekick and flip the toggle to live transcription. While the meeting runs, you can read along with what your attendees are saying.
• Action Items: Jamie provides automatic action items, depending on the type of meeting, which means you do not need to assign them yourself. Once assigned, there is a checkbox that lets you mark when the task is complete. Each task can be prioritised and given a due date.
• Scratchpad: Jamie's Scratchpad gives you a place to write down any thought of your own during the meeting. The Scratchpad lives inside the Sidekick and sits adjacent to the Live Transcript. Anything you add to it does not sync to the meeting summary and stays private to you.

Jamie Gives You 10 Free Meetings A Month With Every Feature Included, And Nothing Locked Behind The Upgrade

Jamie provides 10 thirty minute meetings per month at no cost. We provide all of the features listed, the notes, the transcript, the action items, speaker identification, 100+ languages and unlimited Ask AI, to users prior to requiring payment. Should you be testing a free transcription software prior to requesting funding from others, these three plan limits are likely to be the ones you encounter:

• Full Free Plan: All of the features listed are available on the free plan. These include speaker identification, action items, tags, calendar integration and unlimited storage. There is no restriction on any of them, they are included with your free account and do not need an upgrade.
• Meeting Length: Meeting length varies per plan. A free meeting is limited to 30 minutes, Plus runs to 2 hours, and Pro, Team and Enterprise run up to 5 hours. Therefore a three hour investment committee is one recording rather than three separate segments.
• Export On Every Plan: You can export your meeting notes and the transcript into other applications such as Notion, Google Docs and OneNote. Exporting is available on the free plan.

Jamie Identifies Each Speaker By Name And Matches That Voice In Your Next Meeting

Jamie identifies speakers by name in both the transcript and the notes generated by Jamie. Jamie also utilises Speaker Memory to match that same voice in later meetings. Users can also opt out of Speaker Memory at either the individual user level or the workspace level, which means a person you previously identified comes back identified.

• Speaker Identification And Memory: The transcript and the notes carry the actual name of each participant rather than Speaker 1 and Speaker 2. Jamie also matches the same voice when that participant turns up in a later call. Both of these are available on the free plan.
• 100+ Languages: You get notes and transcripts in multiple languages, including calls recorded partly in one language and then partly in another. No prior notice is required about which languages will be spoken. 100+ languages are supported on the free plan.
• Editable Everything: You can modify your meeting summary, tasks and transcript using the usual editing tools, adding headings, creating lists, adding links, and find and replace. If Jamie got a new client's name wrong, you correct it once after the first call.

Jamie Deletes The Meeting Audio After Transcription And Never Trains Any Model On Your Recordings

Jamie permanently deletes meeting recordings once the associated transcription process is complete. Neither we nor our third-party model providers utilise your meeting data for training purposes. Our security page clearly states this in plain language. If your buyer's legal team wants to verify where Jamie sends the recording, they need to read one sentence on our website. Otherwise, someone must individually contact and verify with each company that we partner with prior to entering into an agreement with us; and meanwhile, your work remains stalled. Here are the three primary questions that they ask:

• The meeting audio gets deleted as soon as the transcript is done, on the paid plans and the free one, and there is no recording left sitting in an account for somebody to dig up a year later.
• No Model Training: Your meeting data is never used to train AI models, not by us and not by our model providers, and this is clearly stated in plain English on our security page. Thusly, you do not need to read a privacy statement to work out how your meeting data is handled.
• AES 256 Encryption: Notes taken during each meeting are encrypted at rest with AES 256, while the connections used to reach them run over HTTPS and TLS 1.2. This makes it possible for your compliance officer to check how your meeting data is protected straight from our security page, without having to send us a questionnaire.

Jamie Stores All Meeting Data In Europe, GDPR Compliant And ISO 27001 Certified

Jamie stores and processes all meeting data within Europe. Jamie is independently audited, GDPR compliant and ISO 27001 certified in our own name. As a result of being DORA-ready, a bank or a fund regulated under DORA can use Jamie. Should your legal or IT department require approval of any transcription tool prior to using it for client work, these three compliance answers are what they will request:

• All of your meeting data is stored and processed inside Europe on every plan, including the free one, and your compliance officer gets that in one line instead of chasing a vendor for proof halfway through a sign-off.
• GDPR Compliant And ISO 27001 Certified: Jamie complies with GDPR and has undergone independent audits to certify that we meet the standards set out in ISO 27001 as a company, not because we use a hosting provider who holds that certificate. Additionally, when a reviewer asks for documentation, you will not spend time explaining why the name on the certificate is a data centre rather than the vendor.
• DORA-Ready: Jamie is DORA-ready, which means banks and funds governed under the Digital Operational Resilience Act may sign off on using Jamie.

Jamie Records From Your Own Machine, With No Bot In Your Client's Attendee List

Jamie records audio directly onto your own device and does not join any of the meetings as a participant. We developed it this way since our typical reader is a guest attending someone else's hosted meeting, a diligence call conducted by the target company or an interview set up by a candidate, and you cannot allow a bot into a tenant you do not control. Nothing appears in the host's attendee list; additionally, there is no bot for their IT department to admit or block. The disclosure comes from you; here are the three ways to do it:

• Verbal Consent:In accordance with our documentation, you remain responsible for telling meeting participants that a recording is being made. Before each call you obtain verbal consent from participants in your own words, and therefore participants are never surprised part way through the call that their conversation is being recorded.
• You can switch on the optional Automated Recording Notice email and it goes to the invitees before the meeting, 4 hours ahead by default, anywhere from 10 minutes to 48 hours if you want to change it. Nothing goes out until you switch it on, your client's contacts never get an email from a tool they have never heard of.
• Meeting Chat Notice: Once you switch this on in settings, Jamie places a notice into the chat window of each meeting run on Microsoft Teams, Google Meet or Zoom from the desktop app, in a place where attendees are already reading messages.

Jamie Records In-Person Meetings, Online Meetings And Files You Upload, Across Five Devices

Jamie records in-person meetings, online meetings held on any platform accessible from your computer, and audio or video files you upload. Since Jamie records the audio on your own device, Jamie supports every online meeting platform installed on your computer, Microsoft Teams, Zoom and Google Meet included, without needing any additional integration or activation. If your week includes meetings held inside a client's office, a video call, and a recording provided to you by someone else, Jamie's three capture methods support all of it:

• In-Person Recording: Using the devices you are already carrying, a laptop or a phone, you can record a conversation in a meeting room on any plan, the free one included. And when a client meeting finishes, you leave their office with a written record of the discussion.
• Every Meeting Platform On Your Device: Jamie records conversations using the audio on the device itself, on macOS or Windows, including Microsoft Teams, Zoom, Google Meet and any other platform running there, without users having to connect an integration before a meeting starts.
• Jamie has apps for macOS, Windows, iOS and Android and the web, and your meetings sync across all of them from one account, so record a client conversation on your iPhone in their office and it is on your laptop by the time you get back.

Jamie Answers Questions Across Your Meetings, Acts In Your Other Tools And Logs To Your CRM

Jamie answers questions across all of the content you have recorded, carries out multi step tasks in the tools you connect to, and logs the meetings you have had into your CRM. We built these three as separate layers so that each one solves a different problem. This way, if you do not need two of them, you do not have to use them. If your transcript is where your work begins rather than where it ends, these three capabilities take the next task off your plate:

• Conversational Intelligence: You can ask Ask AI about any part of your recorded calls, including one question that reaches every call you have recorded with Jamie, summarise many calls at once, and search by tag or by date, with no limit on any plan, the free one included.
• Jamie is agentic, the Agent carries out multi step jobs across the tools you have connected, Agent Connectors authorise it with your own account and Agent Skills hold the instructions you write once. That is the twenty minutes of admin after every call gone.
• Plug And Play Integrations: Jamie logs meetings to HubSpot, Salesforce, Attio and Asana, sends note content into Notion, Google Docs, OneNote and Slack, and offers MCP, webhooks and an API for pulling the data into your own workflow. Because you are not manually typing meeting data into a record somebody else is waiting on, the work of writing up a meeting drops away.

Jamie Pricing

• Free: €0 per month
• Plus: €25 per month
• Pro: €47 per month
• Team: €39 per seat per month
• Enterprise: Custom pricing

Source: meetjamie.ai/pricing, read with the monthly toggle on, 3rd September 2026.

What Jamie users love

Bot free capture, with no awkward bot on the call

"I like Jamie because it's a bot-free AI note-taker that can take notes without an awkward bot on the call." Verified User in Design, G2

Accurate transcription and well organised AI summaries

"The transcription is also usually very accurate. I also like the AI summaries it generates after the call - they're well laid out and organized, especially compared to other AI note-takers." Verified User in Design, G2

A lifesaver for multilingual meetings, online or in person

"Finally an app that... Understands Hebrew. Jaime is also a life saver for: Multi lingual conversations [and] Offline and online meetings." Rachel C., G2

What Jamie users flag

Audio capture needed a little microphone setup early on, since fixed

"I had an issue before with recording the meeting notes because my microphone or something was not set up properly. But the latest update now grabs the meeting sounds with no extra configuration." Verified User in Design, G2

Speaker identification is not infallible, and a good microphone gets the best results

"The speaker identification feature of Jamie, though very useful, isn't entirely reliable or infallible. In our experience, to get the most effective results from Jamie, it was necessary to invest in a good quality microphone." Tom W., G2

A few rough edges and missing integrations, because Jamie is young

"There a few buggy things and missing integrations but Jaime is young and I have full confidence in what their future looks like. It's better than every other app I've tried." Rachel C., G2

2. Happy Scribe: Best Free Transcription Software for Human Proofreading

Best for: Happy Scribe is the free transcription software for people who want an AI transcript in 150+ languages that a real person can then check, recorded on Google Meet, Zoom or Teams.

Similar to: Rev, Sonix, Trint and Amberscript.

Source: HappyScribe

Happy Scribe is a Spanish based company providing three things: an AI generated transcript, a human generated transcript of the same file, and a choice between a notetaker that attends your meeting and a browser based recorder without bots. Their free plan records unlimited meetings on Google Meet, Zoom and Microsoft Teams, and limits each recording to 45 minutes. A file you upload gets a 10 minute trial of AI transcription, after which minutes have to be purchased, and automatic speaker detection is included on every plan.

Key Features

• Human Proofreading: A Real Person Checks The Machine's Work. Happy Scribe sells human proofreading alongside the AI transcript, and their Business plan takes 5 percent off the cost of it. You get a second pass on a recording that has to be right.
• 60+ Languages On The Free Plan, 150+ Across The Product. Happy Scribe's plan comparison prints 150+ languages for their AI speech to text in the free column as well as the paid ones. The same comparison page prints 60+ languages and dialects on the free plan card, which means their own two pages disagree.
• Automatic Speaker Detection: Included On The Free Plan. Happy Scribe identifies automatic speaker detection as a key feature of the free plan. You do not pay to find out which speaker was speaking.
• Action Notes: Follow Ups And Deadlines Pulled Out Automatically. Action Notes in Happy Scribe highlight follow ups, tasks and deadlines automatically after a meeting. Their pages name no owner field on an action item, which means the list arrives with nobody assigned responsibility.
• AI Chat: Three Files A Month On Free, Unlimited On Pro. AI Chat in Happy Scribe spans entire folders and several meetings at one time, which means you do not load one file at a time. The free plan caps it at 3 files a month.

Happy Scribe Holds SOC 2 Type 2, Keeps Its Data Centre In The European Union And Trains Only On Opt-In

Happy Scribe's security page says they are SOC 2 Type 2 certified and describes their data centre as located in the European Union, a Tier IV, PCI DSS and ISO 27001 compliant facility. Those certificates belong to the physical location rather than to Happy Scribe itself. The company further indicates that customer content trains machine learning models only if the customer opts in at signup, that the content is de-identified first, and that permission can be removed at any time from account preferences. Once a customer deletes a file, Happy Scribe retains it for 10 days as a recovery window, and after that the file is permanently deleted.

Pricing

• Free: $0 per month
• Basic: $17 per month
• Pro: $29 per month
• Business: $89 per month

Source: happyscribe.com/pricing, read with the monthly toggle on, 3rd September 2026.

What Happy Scribe users love

A custom glossary keeps oddly spelled names right every time

"I've been using HappyScribe for three years, and I still prefer it over any other A.I. transcription service or notetaker I've tried. I set up a custom glossary for oddly spelled names and words, so my transcripts automatically come out with the correct spelling." Laura Z., G2

The most accurate option this reviewer found for video transcripts

"I love HappyScribe's accuracy. I primarily use HappyScribe for YouTube and other video transcripts, and it's consistently the most accurate option I've found compared to AI LLMs." Tre' H., G2

Speaker detection and timestamps are reliable, and the export is clean

"Fast, accurate transcripts with easy export. HappyScribe's AI transcription is fast, accurate, and easy to use with minimal training. Speaker detection and timestamps are reliable." Monique B., G2

What Happy Scribe users flag

Human accuracy is charged on top of the per transcript price

"I dislike the pricing of the service. I don't like that you charge extra for human accuracy and then also charge per transcript." Tre' H., G2

The human powered transcription and translation is expensive

"The Human-powered transcription and translation is expensive. Sometimes the quality is a little lower than expected, but HappyScribe always makes it right in the end." Torrey M., G2

The style guide will not accept an uploaded document, and speaker names auto-fill with first names only

"The style guide feature could use improvement. It can't accept an uploaded style guide document, so maintaining brand consistency requires manual workarounds. Speaker names auto-fill with first names only." Monique B., G2

3. oTranscribe: Best Free Transcription Software for Typing Interviews Yourself

Best for: oTranscribe is the free transcription software for journalists and researchers who would rather type an interview themselves than upload it, with the audio never leaving their own computer.

Similar to: Express Scribe, InqScribe and Transcribe by Wreally.

Source: Otranscribe

oTranscribe is a free web browser application developed by Elliot Bentley. It provides an audio and video player, a text box, interactive timestamps and keyboard shortcuts on one screen, and it accepts local files as well as web addresses and YouTube links. It does not transcribe anything for you. It removes the part of manual transcription that actually hurts, which is switching between a media player and a word processor forty times an hour. The interface ships in two dozen languages, and it is a project of the MuckRock Foundation.

Key Features

• Keyboard Shortcuts: Pause And Rewind Without Leaving The Text Box. oTranscribe lets you pause with escape, rewind with F1 and fast forward with F2, plus control playback speed using F3 and F4. Your hands stay on the keyboard, which is where most of the lost time goes.
• Interactive Timestamps: Click A Timestamp To Jump Back To That Moment. oTranscribe adds timestamps to the transcript that you click to jump to that point in the audio. Instead of scrolling through the waveform searching for a quote, you click the timestamp.
• Automatic Local Saving: The Transcript Saves Every Second. Every second, oTranscribe saves your transcript to your own browser storage. Closing a tab does not cost you an afternoon of work.
• Open Source Under MIT: The Code Is Public And The Project Is Funded By Donations. oTranscribe is open source under the MIT licence and maintained by the MuckRock Foundation. There is no corporate owner here, which means nobody can buy it and shut it down next quarter.

oTranscribe Keeps Your Audio And Your Transcript On Your Own Computer, And Holds Only Your Account Details

oTranscribe's own page states that your audio file and transcript never leave your computer, and that the transcript is saved to your browser storage every second. No recording is uploaded, which leaves no certificate to request and no retention period to negotiate for your audio. Their privacy policy does say the account details oTranscribe holds sit inside MuckRock's own app and are kept indefinitely. There is also no backup: the transcript lives in one browser on one machine until you export it to Markdown, plain text, .otr or Google Drive.

Pricing

oTranscribe has no cost, no account is necessary, and they do not offer a paid version of oTranscribe. Their website names no cap on how much you transcribe.

What oTranscribe users get

The audio and the transcript never leave your computer

As stated on the oTranscribe website, the audio file you upload and its transcription remain on your computer, and save to browser storage every second. None of this information is ever sent anywhere else. Thus, there is nothing to request certification for, and there are no retention periods to discuss.

Free at any volume, with no account

oTranscribe also requires no account creation, nor do they offer paid versions. The oTranscribe web application offers its interface in twenty-four different languages, and its service worker caches the app when you go offline, allowing you to continue working on your transcription. According to their own webpage, oTranscribe is only compatible with desktop devices.

Open source under MIT, run by the MuckRock Foundation

For this reason, there is no company here that could eventually be bought and closed down next quarter.

What oTranscribe does not do, against the other five tools

It does not transcribe anything for you

oTranscribe does not have speech recognition. Every week you will spend hours typing in your recordings instead of spending those hours doing something else. All five of the other tools listed above will create the transcription for you.

No speaker identification, no summary, no action items

Of the other five tools listed above, four identify each speaker by name. When you are using oTranscribe, you will have to enter the names of each speaker manually, and then you will have to summarize the notes so they make sense to someone who was not present during the meeting, plus extract the action items yourself.

No recorder, and no backup

oTranscribe does not take part in a live meeting, which means you will need to use another app to record those meetings and then import the resulting audio files into oTranscribe. Also, your completed transcripts sit in your local browser storage on one particular computer. oTranscribe does save a backup of your transcript every five minutes and lets you restore it, but warns that backups are not stored for very long. So if you clear your browsing data, and have not exported your transcript before doing so, all of them will disappear.

Source: otranscribe.com, read 3rd September 2026.

4. Otter.ai: Best Free Transcription Software for Live Meeting Notes

Best for: Otter.ai is the free transcription software for people who want a recorder that joins Zoom, Microsoft Teams and Google Meet with speaker identification and AI Chat across their meetings.

Similar to: Fireflies.ai, Fathom, MeetGeek and tl;dv.

Source: Otter

Otter.ai is a recording system that takes down the meeting notes while the call runs. Their Notetaker joins a Zoom, Microsoft Teams or Google Meet meeting as a named participant. Separately, their desktop application for macOS and Windows records the same meetings without a bot, which their help centre calls botless recording. The Basic version provides 300 transcription minutes a month with live transcription, speaker identification and AI Chat. Each conversation is limited to 30 minutes, and their pricing page allows 3 lifetime audio or video file imports.

Key Features

• AI Chat Across Meetings: Included On The Basic Free Plan. On their pricing page, Otter.ai prints AI Chat within and across meetings on Basic. Their help centre caps the free tier at 20 queries a month and 3 per conversation.
• Speaker Identification And Memory: Otter Learns A Voice You Tag. Otter.ai offers speaker identification on Basic, and their help centre says that once you tag a speaker, that speaker is identified in your future meetings. You tag a person once rather than every time.
• Botless Recording: The Desktop App Records Without Joining. Otter.ai's desktop applications for macOS and Windows record Zoom, Microsoft Teams and Google Meet without the Notetaker joining as an attendee. You decide per meeting whether anything joins at all.
• Salesforce, HubSpot And Zapier: On The Pro Plan. According to Otter.ai's pricing page, Salesforce, HubSpot and Zapier integrations come with Pro. Their footnote limits that sync to 1 user on Pro and 5 users on Business.

Otter.ai Is SOC 2 Type 2 Attested, Stores Data On AWS In The United States, And Trains Its Own Models On De-Identified Recordings

According to Otter.ai's privacy and security page, they are SOC 2 Type 2 attested, follow HIPAA requirements, and have built GDPR and CCPA standards into their policy. They are also certified under the EU-U.S. Data Privacy Framework, its UK extension and the Swiss-U.S. framework, with Standard Contractual Clauses. Their own page says their security policies are created based on the ISO 27001/2 framework rather than certified against it. Their privacy policy names Amazon Web Services in the United States, and no European region. On training, they say no customer data trains their AI service providers' models, and that they de-identify user data before training their own. A conversation you delete sits in the trash for 30 days and is then deleted automatically.

Pricing

• Basic: $0 per month
• Pro: $16.99 per user per month
• Business: $30 per user per month
• Enterprise: Custom pricing

Source: otter.ai/pricing, read with the monthly toggle on, 3rd September 2026.

What Otter.ai users love

Staying focused on the prospect instead of writing during a discovery call

"AI Meeting Notes has changed the way I run discovery calls because I can stay focused on how the prospect answers, instead of scrambling to capture budget, competitors, dates, technical concerns, and names while the conversation keeps going." Rostyslav M., G2

Real time transcription with concise automated summaries

"What I like best about Otter.ai is how seamlessly it handles real-time transcription and generates concise automated summaries. It allows me to stay fully engaged during client calls and meetings." Kayreal P., G2

Searchable transcripts and action items you can come back to

"Otterai makes it much easier to keep track of meetings without taking notes manually. I can review transcripts later, search for important discussions, and check action items whenever I need them." Muzammil M., G2

What Otter.ai users flag

Reviewers say language support feels almost English only on multilingual meetings

"I find Otter.ai has limitations with language support, as it operates almost exclusively in English, which makes it struggle with non-English conversations, code-switching, or multilingual team meetings." Verified user, G2

Accuracy drops on noisy audio, several speakers, or strong accents

"What I dislike most is that the transcription accuracy drops when dealing with heavy background noise, multiple speakers talking over each other, or strong accents. It often requires manual editing afterward." Kayreal P., G2

The spread into agents and workflows turns a notes tool into something to govern

"Otter is expanding quickly into agents, workflows, MCP, and more connected sources. That opens up a lot of possibilities, but it can also turn what used to be a simple meeting-notes tool into another system that needs real governance." Rostyslav M., G2

5. TurboScribe: Best Free Transcription Software for Uploaded Files

Best for: TurboScribe is the free transcription software for people working from files they already have, with speaker recognition, 98+ languages and DOCX, PDF, TXT, SRT and VTT exports on the free plan.

Similar to: Sonix, Notta, Descript and Rev.

TurboScribe transcribes audio and video files by upload, by pasting a link including YouTube, or by recording from your microphone. Their free plan limits you to 3 files a day at 30 minutes each, and free files are processed after paying ones. All three accuracy modes are on the free plan, and transcription covers 98+ languages.

Key Features:

• Speaker Recognition On Free: Who Said What, Without Paying. TurboScribe's own feature page describes Speaker Recognition labelling each voice, and it is on the free plan. A two person interview comes back readable rather than as one wall of text.
• 98+ Languages Transcribed, 134+ Translated. Alongside transcription in 98+ languages, TurboScribe also translates a finished transcript into 134+. Whether the recording was made in Spanish, Chinese, French or Japanese, it can be turned into the language your colleague reads.
• 30 Minute Free Uploads, Three A Day. TurboScribe's free plan allows 3 transcripts a day at 30 minutes each. That covers a week of short interviews, and a three hour workshop does not fit inside it.
• 10 Hour Uploads On The Paid Plan. TurboScribe Unlimited lifts each file to 10 hours or 5 GB and lets you upload 50 files at once, individually or in batches. Instead of one file at a time, a back catalogue is cleared in one afternoon.

TurboScribe Encrypts Your Files, Stores Them In The United States And Keeps Them While Your Account Is Active

According to TurboScribe's own pages, your uploaded files, your transcripts and your account details are encrypted both at rest and in transit, and their site names AES-256 encryption. Their privacy page says that if you are in the EU or the UK, information you provide is transferred to and stored in the United States. Their security and privacy FAQ says they may keep your content while your account is active, and their terms say deleted content is permanently removed within 7 days. Neither their privacy page nor their terms names an ISO 27001, SOC 2 or GDPR certification, and their FAQ says they do not train AI or machine learning models on your files.

Pricing

• TurboScribe Free: $0 per month
• TurboScribe Unlimited: $20 per month

Source: turboscribe.ai/pricing, read with the monthly toggle on, 31 August 2026.

What TurboScribe users love

A simple interface and quick uploads

"Turboscribe has made it much easier to turn recorded conversations and videos into editable text without spending a lot of time typing everything manually. The interface is simple, uploads are quick." Muzammil M., G2

It removes the manual typing from a working day

"TurboScribe makes it much easier to turn audio content into written text without spending a lot of time on manual typing." Ishan S., G2

Convenient for basic transcription work

"The interface is simple, and the idea of quickly transcribing audio and video files is useful. The service can be convenient for basic transcription tasks when it works as expected." Marat A., G2

What TurboScribe users flag

The paid Unlimited plan met upload limits after a bulk upload

"The biggest issue is the lack of transparency around the Unlimited plan. I paid for TurboScribe Unlimited because it was presented as Unlimited Transcriptions. After payment and bulk upload, I ran into upload/transcription limits." Marat A., G2

Background noise or several speakers means checking the transcript afterwards

"The results are very good when the audio is clear, but recordings with background noise or multiple people talking at the same time sometimes need a quick review." Muzammil M., G2

6. OpenAI Whisper: Best Free Transcription Software for Unlimited Local Transcription

Best for: OpenAI Whisper is the free transcription software for people who want unlimited transcription on their own machine, covering 98 languages besides English, with nothing uploaded anywhere.

Similar to: MacWhisper, whisper.cpp, Buzz and Meetily.

OpenAI Whisper is a speech recognition model rather than a product with a login. OpenAI released the code and the model weights together under the MIT licence, which means you install it and run it from your own command line, at any volume, for nothing. The same command translates spoken words into English and identifies the language of a recording as well as transcribing it. There is no monthly ceiling of any kind, and it is the one tool here that asks you to open a terminal before you get a single word of text.

Key Features:

• MIT Licence: Free At Any Volume, Forever. OpenAI Whisper is open source software released under the MIT licence, which means there is no fee for installing it or for running it. Whether you have one hour of archive or a thousand hours, it costs you nothing to process.
• 98 Languages Besides English: Covered By The Training Data. According to the model card, the non-English training data represents 98 different languages, and OpenAI notes that accuracy varies widely between them. A recording in a smaller language has a real chance of coming back usable.
• Six Model Sizes: You Trade Accuracy Against Your Own Hardware. OpenAI Whisper ships models from tiny at 39 million parameters to large at 1.55 billion, plus a turbo model, needing roughly 1 GB to 10 GB of video memory. An older laptop still runs the small model, and the command falls back to your processor when there is no graphics card.
• Local Processing: The Model Runs On The Machine You Installed It On. OpenAI Whisper runs from your own command line, and downloads the model weights once on first use. A recording covered by an NDA never travels to a vendor you would otherwise have to vet.

OpenAI Whisper Runs On Your Own Machine, And The Recording And The Transcript Never Reach A Vendor

Transcripts created with OpenAI Whisper are saved in txt, vtt, srt, tsv, json or jsonl format, or in all formats at once, to the same directory where you ran the command.

Pricing

OpenAI Whisper is free and open source under the MIT licence. Furthermore, there is no paid option.

Source: github.com/openai/whisper, read 3rd September 2026.

What OpenAI Whisper users ask for

Two of the most popular user questions in OpenAI's repository are about gaps. They are user question threads rather than reviews, but both reference a gap.

Speaker diarization is the most asked for missing piece

"Is there a simple way to do speaker diarization with whisper ?" dehma01, GitHub

Live streaming is the second

"Is it possible to use whisper for streaming tasks (with syntax)? For example, would it be possible for whisper to be bound to a websocket of streaming PCM data packets?" davidhariri, GitHub

What OpenAI Whisper does not do, against the other five tools

No speaker labels come back

Neither the OpenAI Whisper README nor its model card includes speaker diarization. Their own Community Discussions thread covers running Whisper alongside a separate diarizer, which is a community thread rather than a shipped feature. If you run the model on its own, everything in the entire transcript will appear as one big block of text. Therefore, you will need to go through the transcript and manually assign which individual is speaking. Jamie, Happy Scribe, Otter.ai and TurboScribe all identify each speaker by name for you.

No recorder, no summary, no action items, no interface

Whisper is called from a command line and comes with a Python library documented in their README. For that reason, if your team is not technical, a member of your team who is capable of writing code will be responsible for setting up the model and writing the code that processes the transcript.

Nothing is stored for you, and nothing syncs

No account required and no cloud storage. Transcripts generated by OpenAI Whisper are saved to the same folder you called the command from. Therefore, you will need to save and back up every single transcript you generate. If the laptop goes, the archive goes with it.

Before You Choose A Free Transcription Software, Ask These Two Questions

While the number of minutes allowed for a free transcription service may seem like the significant limit to consider, the true determining factor is really the total amount of time allowed for one recording, the job you need done, and how the other people on the call are being identified.

The other three factors are the pre-installed free tools already available on your device, the accuracy level of the generated transcript, and the file formats used to go in and out.

Are Transcribing A Recording And Transcribing A Live Meeting Two Different Jobs?

Ultimately, there are two primary forms of free transcription software. They include software that converts existing files into text, and software that generates text from a live meeting as it happens. Very few pieces of free software fall within both categories. Choosing the incorrect category is the main cause of why most free plans fail.

If you want to use a recording you already possess, meaning an interview from your phone, a downloaded webinar or a voice memo, you will want an uploader. OpenAI Whisper produces a transcript of your recordings on your local machine at any length. TurboScribe allows you to upload 3 files daily at no greater than 30 minutes in length. Finally, oTranscribe includes a playback component as well as a text box where you manually create your own transcript.

Conversely, if you are seeking to document a conversation that has yet to occur, you will want a recorder. Jamie uses your laptop or phone to record meetings and does not join meetings as a participant. Otter.ai and Happy Scribe each offer several alternatives, including a notetaker that joins the meeting on Zoom, Microsoft Teams or Google Meet, or a bot free recorder that does not actively participate in meetings.

If you require both, meaning previously recorded files and newly conducted meetings, Jamie and Happy Scribe both enable you to upload previous recordings and record new meetings. Therefore, one tool covers both last week's interview and Thursday's client call.

Does Free Transcription Software Need Consent, And Does Every Tool Handle Consent The Same Way?

You make a decision about somebody else when you decide whether to record a conversation, and none of the six tools listed above will make it for you. What differs is the way each of the tools presents itself to the other participants on the call.

Otter.ai and Happy Scribe provide both options. Each includes a notetaker that attends the meeting and is listed among the attendees, which represents a type of disclosure, and each is the thing your client's IT department can block. Alongside its bot-based service, Happy Scribe offers a browser-based free recorder. Otter.ai provides a desktop recorder for Mac and Windows that does not use a bot.

Jamie captures meetings using your own device and does not participate in the meeting itself; therefore, you will have to disclose that Jamie is capturing the meeting, and Jamie gives you three ways to disclose it. You may obtain verbal consent in your own words before the meeting. Jamie sends an optional Automated Recording Notice email to all invitees before the meeting, four hours in advance by default and configurable from ten minutes to forty eight hours, and no emails are sent unless you switch it on. The third option is the meeting chat notice that Jamie posts inside calls on Microsoft Teams, Google Meet and Zoom from the desktop app, once you enable it in settings.

TurboScribe, oTranscribe and OpenAI Whisper do not interact with live meetings at all, which means whatever was used to capture the original audio established consent, and that was you.

Read more on bot free AI note takers and how the disclosure works without a participant in the call.

Final Verdict on What's the Best 6 Free Transcription Software?

The best free transcription software depends entirely on what you need right now; if it's a bot free recording of your own meetings, with the notes, the transcript and the speaker names on the free plan, then, Jamie can be the best choice among the free transcription software. Let's do a quick recap of why:

• Privacy And Preserving The Natural Conversation: Jamie records the audio on your own device and does not join the meeting, which means nothing appears in your client's attendee list. The disclosure comes from you, through verbal consent in your own words, the optional Automated Recording Notice email to the invitees, and the meeting chat notice.
• Advanced Technology Used For Speaker Identification: Jamie identifies each speaker by name and matches those voices in later meetings. Tested against 4 other AI note takers, Jamie scored 98 out of 100 for identifying speakers.
• Rooted Privacy Laws And Highest Security: Jamie stores and processes all meeting data in Europe, and is GDPR compliant and ISO 27001 certified in our own company name. Jamie is also DORA-ready.
• Values Your Privacy And Keeps AI Training Out Of It: Jamie deletes the meeting audio after completing the transcription, and does not use your data to train our AI models or anybody else's.
• A Free Plan That Is The Whole Product: On Jamie's free plan you get 10 meetings a month at 30 minutes each, with every feature included. Once a meeting is done you can export the notes and the transcript.

You can also pick a different tool when the job is different.

• Choose Happy Scribe when a recording has to be right and you want a real person to proofread the AI transcript, in one of the 150+ languages their own pages name.
• Choose oTranscribe when you would rather type an interview yourself and you need the audio to never leave your computer, with no account and no upload.
• Choose Otter.ai when your work happens in Zoom, Microsoft Teams and Google Meet and you want a recorder that joins the call and answers questions across your meetings.
• Choose TurboScribe when you are working from files you already have and you want speaker recognition and 98+ languages on a free plan.
• Choose OpenAI Whisper when you have a large archive, a machine to run it on and no budget at all, and you can live without speaker labels.

For everything else, Jamie records from the device you already carry with no bot in anybody's call; Jamie generates the notes, the transcript, the action items and the speaker names on every plan including free; and Jamie keeps all of it in Europe, GDPR compliant and ISO 27001 certified, with the meeting audio deleted once the transcription is finished.

‍

FAQS: Best Free Transcription Software

What is the best free transcription software?

The free transcription tool that performs best at providing bot free meeting notes that include the transcript, the action items and the speaker names is Jamie. If you want an AI created transcript that can be reviewed by a human, use Happy Scribe. If you prefer to type an interview yourself but still need to keep the recording locally on your machine, then try oTranscribe. For those who wish to capture live recordings during meetings with a recorder that joins Zoom, Microsoft Teams or Google Meet, Otter.ai would be a good option. To upload your files and use speaker recognition, consider TurboScribe. And if you want unlimited local transcription, OpenAI Whisper should meet your requirements, given that it falls under the MIT licence.

Which free transcription software works without a bot joining the meeting?

Jamie records the audio to your device and does not connect to the meeting itself as a participant. Therefore, there are no entries added to your client's attendee list, nor are there bots for their IT team to admit or block. Happy Scribe and Otter.ai both offer bot free recording, through a browser recorder and a desktop app respectively. What Jamie provides on top of that is the disclosure: verbal consent in your own words before the call starts, the optional Automated Recording Notice email before the meeting, and the meeting chat notice once you enable it.

What is the catch with free transcription software?

The problem here is the limitations below the monthly limit. Otter.ai has a free tier of 300 minutes per month, but there is a cap of 30 minutes per conversation and there can be only three lifetime file imports. Meanwhile, TurboScribe provides three free transcripts a day and each file has a limit of 30 minutes. Happy Scribe allows as many meetings as you want, but each meeting is limited to 45 minutes, and once the audio is a file you upload, you get a 10 minute limitation. Finally, Jamie limits the number of meetings you can have to ten per month, each at 30 minutes long, and all features are included on that plan.

Does free transcription software train AI on my recordings?

Depending upon the vendor, the extent of that vendor's ability to use your data for model training is found on the vendor's own website. Jamie never uses your data for model training, and neither do our model providers. TurboScribe says the same in their Security FAQ, that they do not use your media files or transcripts to train any AI or machine learning models. Similarly, Otter.ai indicates that customer data is not used to train their AI service providers' models, and that they de-identify user data before using it to train their own models. Only when you explicitly opt in during signup for Happy Scribe is data used to train, and Happy Scribe allows you to withdraw your consent at any time thereafter. Both oTranscribe and OpenAI Whisper never receive your audio files at all.

Where is my meeting data stored with free transcription software?

Jamie stores and processes all meeting data within Europe. Jamie is both GDPR compliant and ISO 27001 certified. According to Happy Scribe's security webpage, their data centre is located in the European Union. In contrast, according to Otter.ai's privacy policy, Amazon Web Services is listed as being located in the United States. TurboScribe's privacy webpage states that information collected from users in the European Union or the United Kingdom may be transferred to and stored in the United States. Neither oTranscribe nor OpenAI Whisper store anything at all, since the audio files never leave your local system.

How can I transcribe audio for free right now?

There are three options for completely free transcription of audio. First, upload a file to the free tier of either TurboScribe, which allows 3 files a day, each capped at 30 minutes, or Happy Scribe, which gives an uploaded file a 10 minute allowance. Second, run it locally with OpenAI Whisper, free in any quantity under the terms of the MIT licence. Third, use Jamie's free plan to record the meeting live, which also includes the notes, the transcript and the speaker names, for up to 10 meetings per month at 30 minutes each.

Is Google Docs Voice Typing free transcription software?

Google Docs Voice Typing is another option for free transcription. While it is indeed free, it is a form of dictation rather than transcription. Specifically, it generates written text from spoken words into a Google document in real time. As such, Voice Typing cannot take an existing audio file, it cannot identify each speaker by name in a recording, and it will not work if you are trying to generate a transcription of a pre-recorded meeting. Thus, if you wish to create a transcription of a recording made by two or more people, as opposed to creating a document while you speak, then you will require one of these six tools.

Read more

• Recording meetings in Europe? Read transcription software in Europe.
• Meetings happening in a room rather than on a screen? Read AI note takers for in-person meetings.
• Leaving Otter.ai? See the shortlist in Otter alternatives.
• Want the notes as well as the transcript? Read transcription software for meetings.
• Comparing the AI transcription category as a whole? Read AI transcription software.