Your First Session
Record your first live transcription in a few minutes, from picking a microphone to reading your transcript.
This guide walks you through your very first live transcription in Kalima, opening the Studio, choosing your audio, pressing record, and turning the result into something you can read, refine, and share. The whole thing takes just a few minutes.
Before you start
You only need two things: a verified Kalima account and a microphone (your laptop's built-in mic is fine to begin with). If you have not created your account yet, start with Create Your Account, verifying your email is what unlocks recording and AI features.
Recording draws from your transcription minutes. Every plan includes a weekly allowance, so you can record your first session right away without buying anything. You can review the details on Plans & Pricing.
Step 1: Open the Studio
The Studio is where you record, watch your transcript form, play it back, and export it. Open it from the Studio link in the app to start a new session.
When you create a new session, Kalima walks you through a short setup with steps for your audio, languages, optional translation, and optional context. You can accept the defaults and start in seconds, or fine-tune each step, it is entirely up to you.
Step 2: Choose your audio source and microphone
Pick where your sound is coming from. For a first session, your microphone is the simplest choice.
Choose Microphone to transcribe your own voice or a conversation in the room. Allow microphone access if your browser prompts you, then select the device you want from the list. A live level meter sits beside each device, so speak a few words and confirm the bars move, that tells you the right microphone is picked up and working.
Pick a quality mode
If you are recording from a microphone, choose the capture mode that fits your room:
- Enhanced is the default. It turns on echo cancellation, noise suppression, and auto-gain, which is what you want for a laptop mic or any room with other people in it.
- Raw records unprocessed audio, with all of that turned off. Use it with a decent microphone in a quiet room, ideally on headphones.
You can tap Test modes to compare the two on a quick sample. Your microphone and quality choice are remembered for next time.
The single biggest factor in accuracy is clean input. A good microphone close to the speaker, in a quiet room, beats almost any other adjustment. See Tips for Accurate Transcription for more.
Step 3 (optional): Add languages and translation
Kalima detects the spoken language automatically, you never have to pick one to get started, and it handles conversations that switch between languages mid-sentence. If you already know which languages will be spoken, you can add them as expected languages in the Languages step to nudge accuracy higher, especially for accents and similar-sounding languages.
Want a live translation too? In the Translation step, choose:
- One-way to translate everything into a single target language.
- Two-way to translate between two languages in both directions at once.
Translated text appears inline alongside your transcript as you speak. Higher plans include enhanced translation quality automatically. Learn more in Live Translation.
Step 4: Press record
When you are ready, press record. Kalima connects, then begins capturing, and your words start appearing on screen.
Press the record button. The session connects and capture begins.
Speak naturally. Text first appears as a fast preliminary result, then settles into the confirmed version a moment later. This brief refinement is normal, confirmed text can trail the live audio by a second or two while the engine finishes each phrase.
Watch the transcript fill in. Different speakers are separated automatically and labeled Speaker 1, Speaker 2, and so on, each in its own color. Per-segment badges show the detected language.
Keep an eye on the latency badge in the header. Green means results are arriving fast and your session is healthy; amber or red points to a slow or unstable connection or a high-latency device.
Press stop when you are finished. Your session is finalized and ready to review and play back.
You can stop and start more than once within the same session. Each take is added to the same transcript on one continuous timeline, so meetings with breaks or multi-part recordings stay together. Use pause if you just need a short break without ending the take.
What if my connection drops mid-recording?
Don't worry, keep recording. Your audio is saved on your device, Kalima reconnects automatically, and any stretch that was missed during the interruption is re-transcribed and merged back into the right place afterward. Recovered sections are clearly marked. You can read the full story in Reliability & Recovery.
Recording is metered against your balance. If you run low, you'll see a warning with your remaining time; if it runs out, a short grace countdown appears and recording stops cleanly so there are no surprises.
Step 5: Explore your transcript
Once you stop, your transcript is yours to read, refine, listen to, and share, all in the Studio.
Try an AI summary
Your transcript becomes far more useful with one click. Open the AI panel, select the Summary tab, and generate a structured summary, it streams in as it is written, with a short TLDR, key takeaways, decisions, action items, topics, and open questions. You can view it as an Overview, a Tasks list, a Topics map, or copyable Markdown. The summary saves automatically, so it is there when you come back.
AI features require a verified email, and the output is written in your account's language. If you skipped verification at sign-up, finish it now to unlock summaries, chat, AutoCorrect+, and read-aloud.
Quick troubleshooting
A couple of things people hit on their first try:
Click "Allow microphone access" and accept the browser prompt. If you previously blocked it, re-enable microphone access for the site in your browser settings and reload the page. The level meter beside each device confirms it is picking up sound. More fixes in Microphone & Audio Issues.
Switch the audio source to Desktop audio, and when your browser asks what to share, pick the right tab, window, or screen and make sure the "share audio" checkbox is enabled. Some browsers only expose tab audio, the desktop app captures system audio natively. See System & Meeting Audio Capture.
That is expected. Kalima shows a fast preliminary result first, then refines it into the confirmed version a moment later. Confirmed text can trail the audio by a second or two, which keeps the transcript clean instead of choppy.
Move closer to the microphone, switch to a quieter spot, check that you are on Enhanced mode, and add expected-language hints if you know the languages. For names and jargon, add session context. See Tips for Accurate Transcription.
What next
You have recorded your first session, here is where to go to get more out of Kalima.