Frequently Asked Questions

Kalima turns spoken conversations into accurate, searchable text in real time. You can transcribe live from your microphone or your computer's audio, upload existing recordings, translate as you go, and use AI to summarize and explore what was said. It works in your browser and as a downloadable desktop app. See What is Kalima.

Yes. The Free plan needs no credit card and includes 120 transcription minutes (2 hours) per week, live translation, and AI summaries. See Plans & Pricing.

At least 8 characters, including one uppercase letter, one lowercase letter, and one number. You can also sign up with Google instead of a password. See Create your account.

You can look around right away, but some features stay locked until you verify, including AI features, creating new sessions, and privacy actions like data export and account deletion. Your free weekly minutes are also withheld until you verify. The verification link is valid for 24 hours. Check your spam folder, and use Resend if it lapsed. See Create your account.

Open the Studio, choose your audio source and microphone, optionally add expected languages and translation, then press record. Text appears as you speak; press stop to finish. See Your first session.

No. Kalima detects the spoken language automatically and handles mid-conversation switches, labelling each part of the transcript. If you already know what will be spoken, adding expected-language hints noticeably improves accuracy. About 60 languages are available in the picker. See Languages & detection.

Live transcription shows a fast preliminary result first, then settles into the confirmed version a moment later. Confirmed text can trail the audio by up to about 2 seconds, which is normal and keeps results from looking choppy. See Live transcription.

Yes. Speakers are separated automatically (Speaker 1, Speaker 2, and so on), each in a distinct color, and you can rename them to real names. Numbering stays consistent even across several start/stop takes in one session. See Speakers & diarization.

Enhanced mode is the default and the right pick for most rooms, since it adds noise suppression, echo cancellation, and auto-gain. Switch to Raw mode when you have a decent microphone in a quiet room and want raw, unprocessed audio. See Audio sources & quality.

Add session context: a short description plus key terms, names, and acronyms. This helps the engine spell specialized vocabulary correctly instead of guessing it phonetically. See Improve accuracy with context and Tips for accurate transcription.

Yes. Each take is added to the same session and the timeline continues, so meetings with breaks or multi-part recordings stay in one transcript. See Live transcription.

Yes. In the Translation step choose one-way (everything into a single target language) or two-way (between two languages in both directions). Translated text appears inline with the live transcript. See Live translation.

Translation is available on all plans. Pro, Team, and Enterprise plans automatically get enhanced AI translation quality with no extra setting to toggle. See Live translation.

No. When you enable translation on a file upload, it is processed alongside transcription at no additional cost. That is the "2-for-1" feature. You are still billed only for the audio's transcribed duration. See Transcribe an uploaded file.

Audio: MP3, WAV, M4A, AAC, FLAC, OGG, Opus, WebM, AIFF, and AMR. Video: MP4, MOV, and WebM (the audio is extracted automatically). WhatsApp voice messages (.opus) are supported. See Supported formats & limits.

Up to 5 hours (300 minutes) per file on every plan. Maximum file size depends on your plan: Free 100 MB, Plus 500 MB, Pro and Team 1 GB, and Enterprise has no fixed size limit. See Supported formats & limits.

Free: 20 uploads, Plus: 50, Pro: 100, Team: 100, Enterprise: unlimited. When you hit the limit you'll be prompted to upgrade. See Supported formats & limits.

No. Once the file finishes uploading, processing continues in the background and you can leave the page. Come back later to find the finished transcript in your sessions. (If you leave while the file is still uploading, the upload is cancelled.) See Processing & costs.

You're charged from your time balance based on the actual transcribed duration. Kalima shows an estimated cost and remaining balance before you start, holds an estimate when you begin, and reconciles to the exact amount when it finishes, refunding any difference. See Processing & costs.

A short TLDR, key takeaways, decisions, action items, topics with key points, open questions, and a participants list. You can view it as an Overview dashboard, a Tasks list, a Topics map, or copyable Markdown. It adapts to the session: meetings highlight decisions and tasks, while a lecture or podcast highlights ideas and takeaways. See AI summaries.

Summaries stay faithful to the transcript and won't invent decisions, owners, dates, numbers, names, or quotes. AI chat answers are grounded only in that session's transcript, and the assistant tells you when the answer isn't there. See AI chat.

Summaries, chat answers, automatic titles, and AutoCorrect+ notes are written in your account's display language, even if the transcript is in another language. See AI features overview.

Use AutoCorrect+ (Mark & Fix): mark one occurrence, type or AI-suggest the correction, and tick "also fix others". The AI checks each other occurrence and only replaces the ones that are genuinely the same mistake. Fixes apply at the word level, so audio and karaoke timing stay aligned. See AutoCorrect+.

AI usage is metered by weekly quotas that reset every Monday (00:00 UTC). For example, AI summaries are Free 10, Plus 20, Pro and Team 75, and unlimited on Enterprise; chat messages are Free 25, Plus 125, Pro and Team 500, and unlimited on Enterprise. See the full table in AI usage & limits.

All AI features share a short-term per-user rate limit. If you fire many requests in quick succession you'll be asked to wait a moment, then you can continue. See AI feature issues.

Included minutes come with your plan and reset every Monday (00:00 UTC). Credits are prepaid transcription time you buy separately and don't reset each week. Recording draws from your included minutes first, then from credits. See Minutes, credits & balance.

The base rate is 1 EUR = 60 minutes. Preset packs: 4.99 = 5 hours, 9.49 = 20 hours, 14.99 = 40 hours, 24.49 = 72 hours (Popular), 49.95 = 155 hours, 99.90 = 333 hours. Larger packs give more hours per euro. See Minutes, credits & balance.

You're warned beforehand. When a limit (transcription minutes, uploads, or AI usage) is reached, that feature pauses until your next weekly reset or until you upgrade. You can also buy prepaid credits to keep recording past your weekly minutes. See Usage & quotas.

Yes. Upgrades take effect immediately with prorated billing (you can preview the charge first); downgrades apply at the end of your current period. You can pause an active plan for 1 to 3 months, or cancel and keep access through the end of the billing period. See Managing your subscription.

Included weekly minutes are withheld until you verify your email (and while an account is under review). Purchased credits remain usable in both cases. See Billing & balance issues.

You pay for each active member (owner, admin, or member). Add or remove seats anytime and billing adjusts automatically. Viewers who join sessions by share link do not count as seats. See Billing & seats.

A workspace is a personal folder just for grouping your own sessions. A project is a folder that also carries shared transcription settings and can be shared with collaborators. An organization is a shared container for a company, with members, teams, and billing. See Organizations overview.

Not necessarily. If your share link has a public grant, anyone with the link can view it. If it requires sign-in or is limited to specific people, organizations, or teams, viewers must be signed in and matched. See Sharing a session.

Yes. In the share dialog you can set a password and/or an expiry date and time, after which the link stops working. You can change or remove either at any time, and disabling a link instantly disconnects anyone watching. See Sharing a session.

Yes. Share a session while you record and viewers see the transcript (and translation) update in real time, with no account needed for a public link. Viewer mode is read-only, but viewers can use AI chat and see a read-only summary. See Live viewer mode.

Organization and project collaboration invitations expire after 7 days. After that, send a fresh one. Invites must be accepted from the same email address they were sent to. See Members & roles.

Creating an organization requires a paid plan: Plus can own 1, Pro up to 3, Team and Enterprise unlimited. Free plans can't create organizations. See Organizations overview.

After 5 failed sign-in attempts, an account is temporarily locked for 15 minutes to protect it. Wait and try again, or reset your password. See Sign-in & account issues.

Select Forgot password on the login page, enter your email, and follow the reset link (valid for 1 hour). If you signed up with Google there's no password to reset. Just use Continue with Google. See Account security.

Open the Account tab in your profile and choose Request data export. You can request once every 7 days, and your data is delivered within roughly 1 to 2 weeks. See Your data & privacy.

Go to the Account tab, then Danger Zone, then Delete account; type DELETE to confirm (and enter your password for email/password accounts). It is permanent and cannot be undone. Your transcripts and audio are removed, your profile is anonymized, and any active subscription is cancelled. See Your data & privacy.

A sign-in session lasts up to 7 days, after which you'll be asked to sign in again. Signing out clears locally cached app data on that device, which is helpful on shared computers. See Account security.

Recording keeps going and your audio is saved locally. Kalima reconnects automatically; anything missed during the outage is re-transcribed afterward and merged into the transcript in the correct place, with recovered stretches marked. You'll see a reconnecting or restoring status while it works. See Reliability & recovery.

Yes. You can record offline, with your audio captured locally and transcription produced automatically once you reconnect or stop, using your session's original settings. See Reliability & recovery.

Both. Search covers session names plus the spoken words in your transcripts and translations, ranks results, highlights the match, and lets you click straight to that moment in the audio. It searches sessions saved to the cloud and needs a connection. See Navigating Kalima.

The website handles core transcription, uploads, AI, and sharing. The desktop app adds things browsers can't: automatic meeting detection with one-click recording, capturing your computer's audio, running in the system tray, global shortcuts, floating picture-in-picture, and local screen recording. Both share the same account and data. See Web app vs. desktop app.

Windows, macOS, and Linux. Windows uses a standard installer, macOS comes as a disk image (.dmg), and Linux as an AppImage or .deb package. See Installing the Desktop App.

Only if you turn it on. With meeting detection enabled it shows a Meeting Assist panel you can accept or dismiss; if you also enable auto-start for high-confidence calls, it begins recording on its own. Meeting detection is available on Plus and above. See Meeting detection & auto-recording.

Yes. The desktop app can capture your computer's meeting audio and mix in your microphone, so both sides are transcribed. On Windows it can isolate a specific app's audio; on other systems it captures system audio (you may be asked to grant a screen/audio sharing permission). See Capturing audio & screen.

Not if you turn on Stealth Mode. With it on, screen shares, recordings, and screenshots show a black rectangle where Kalima would be, so your live transcript stays private to you (most reliable on Windows). See Privacy & stealth mode.