Speakly

What is Speakly?

Speakly is Genspark's AI voice input app that lets you replace typing with speech in any application on your computer or phone — up to 4x faster than the keyboard.

Core capabilities:

Voice-to-text — speak and it types, in real time, across 100+ apps including email, docs, Slack, CRM, and more
AI auto-polish — automatically removes filler words like "um" and "uh", fixes misspeaks, and cleans up formatting
Custom instructions — set up dedicated modes like translation, professional rewrite, or terminal commands and switch with one keypress
Genspark Agent Mode — double-press to activate, directly invoke deep research, AI Slides, AI Sheets, and other powerful capabilities
Multi-language support — 100+ languages with automatic detection, no manual switching

Quick Start

The following walkthrough uses the desktop client (Mac / Windows) as an example. On mobile, no shortcut setup is needed — just download, install, log in, and tap the in-app button to activate voice input in any text field.

Download and install Speakly

Visit speakly.ai and download the client for your platform. Available on Mac, Windows, iPhone, and Android.
Sign in with your Genspark account

Sign in with your existing Genspark account — Speakly is integrated with Genspark, so no separate registration is required.
Grant microphone and accessibility permissions

On first launch, you'll be prompted to grant microphone access and accessibility permissions (used to insert text into any app). Follow the prompts to complete authorization. On Mac, you can verify under System Settings → Privacy & Security → Accessibility.
Confirm your shortcut key

The onboarding flow will prompt you to press the default shortcut to test it (the button turns blue when it triggers correctly). You can change the shortcut here, or adjust it later anytime in the Shortcuts page.
Start using it

Place your cursor in any text input field, press and hold the shortcut to start speaking, release when done — the text is automatically inserted at the cursor position. You don't need to choose your words carefully; the AI will clean it up into polished text.

Feature Guide

Default Shortcuts: Three Modes

Open Shortcuts in the left sidebar. The top Default Shortcuts section configures the three core mode shortcuts:

Dictation (press-and-hold) Press and hold the shortcut to speak; release to insert. Suitable for most scenarios — the AI automatically cleans up filler words, misspeaks, and paragraph structure.

Hands-free Mode Press the shortcut once to start recording, press again to stop and insert — no need to hold down the key throughout. Ideal for long-form dictation.

Agent Mode Double-press the shortcut to summon the Genspark Agent dialog. Speak your request and let Genspark handle it on the spot. Best for quick, in-the-flow tasks where you don't want to switch to a browser.

For example:

Reading a tweet or web article: "Help me understand what this is saying"
Writing an email: "Polish this paragraph to sound more professional"
An ad-hoc question: "Quickly look up what XXX means"
An incoming English email: "Summarize this in bullet points for me"
Reading technical docs: "Explain this paragraph in simpler terms"

Shortcut notes:

Dictation and Agent Mode share the same shortcut key (Mac default: fn; Windows default: Right Alt). Press-and-hold triggers Dictation; double-press triggers Agent Mode. Changing one updates the other.
Hands-free Mode uses an independent shortcut. No default is assigned — set one yourself as needed.
All three modes can be toggled independently on the Shortcuts page.

Selection-based Voice Commands

An advanced Speakly workflow: select some text first, then press and hold the shortcut and speak your instruction. The AI processes the selected content according to what you say.

How to use:

Select a piece of text in any application
Press and hold the shortcut, then speak your instruction
Release — the text is replaced according to your instruction

Typical use cases:

Translate / Rewrite:

Select Chinese text: "Translate to English"
Select casual phrasing: "Make it more professional"
Select a paragraph: "Rewrite in a friendlier tone"

Summarize / Understand:

Select a long article: "Tell me what this article is about"
Select an email: "Summarize into three key points"
Select technical docs: "Explain this in plain language"

Format cleanup:

Select a list: "Format this as an email"
Select meeting notes: "Organize into structured meeting minutes"
Select a draft: "Rewrite as a short Slack-style message"

Chained instructions are supported. For example, "First polish, then translate to English" — Speakly will execute the two steps in order.

Custom Instructions (Shortcuts)

Custom instructions let you define a fixed transformation rule for a specific scenario, bound to its own shortcut key. Press the key and speak — the rule fires automatically without needing to repeat the instruction.

Where to set them up:

Click Shortcuts in the left sidebar. The page has two sections:

Default Shortcuts (top): manage Dictation / Agent Mode / Hands-free Mode shortcuts
Recommended Shortcuts (bottom): built-in custom instruction templates, toggle on to use

Enable a built-in template:

In the Recommended Shortcuts area, find the template you want, toggle the switch on, confirm its hotkey (e.g. Left Opt + Z), then press-and-hold that combination and speak to trigger it.

Add your own custom instruction:

Click + Add in the top-right of the page
Fill in the name (used to identify the instruction)
Fill in the AI Prompt — telling the AI how to process what you say each time
Set a unique shortcut key combination
Save it. Press-and-hold the shortcut and speak to trigger it.

Built-in instruction templates (toggle on to use):

Translate to English — speak in any language, automatically translates to fluent English
Terminal Helper — describe operations in natural language, automatically converts into runnable terminal commands
Professional Rewrite — turns casual speech into polished, professional written language
Chaos Mode — adds emojis and memes, ideal for casual contexts
Buzzword Mode — converts simple ideas into corporate-speak full of industry jargon

Custom examples:

"Format what I say as Notion-style bullet points" "Output everything in Slack message style" "Organize my thoughts into meeting-minutes format"

Personal Dictionary

If your work involves proper nouns, brand names, people's names, or industry-specific terms, add them to Speakly's personal Dictionary so the AI can learn them ahead of time and avoid recurring transcription errors.

Where to access it: the Dictionary page in the left sidebar of the Speakly client.

Two ways to add words:

Manual add — open the Dictionary page → switch to the Manually Added tab → click Add and enter the word you want preserved
Auto add — when you correct a word in a transcription, Speakly will offer to save the corrected spelling to your Dictionary; the next time you say the word, Speakly will output your saved version

Capacity: up to 8,000 entries.

Cross-device sync: the Dictionary is tied to your Genspark account and syncs across all devices when you're signed in.

Undo auto-add: if an automatically added entry isn't right, simply delete it from the Dictionary page.

Meeting Notes

Beyond real-time dictation, Speakly can record full meetings end-to-end — capturing audio, transcribing, and generating AI summaries. Click Meeting Notes in the left sidebar; the top of the page offers three entry points:

1. New Note (record and save a meeting)

Click to start recording the current meeting or conversation. Speakly records and transcribes in real time and, when you stop, produces a full transcript and AI summary saved to your Notes list. It recognizes multiple speakers and outputs speaker labels with timestamps.

While recording, you can also open the Floating Subtitle window to follow along, pick a target language to see both the original and a translation, and switch the audio source between microphone and system audio — these share the same real-time engine as Live Translation below.

2. Live Translation (real-time transcription + translation)

If you just want to follow along with live subtitles and don't need to save a Note, Live Translation is the lighter option. Click to launch a Floating Subtitle window pinned on top of any app showing real-time speech-to-text.

Pick a target language (Chinese, English, Japanese, Korean, Spanish, Portuguese, French, German, Italian, Russian, Arabic, and more — 15+ supported) and Speakly will display both the original transcript and a real-time translation. You can independently toggle the original and the translation — show original only, translation only, or both side-by-side. The audio source also switches between microphone and system audio, making it useful for cross-language meetings, foreign-language lectures, or watching foreign-language video.

3. Import File (upload audio / video for transcription)

Already have a recording? Upload it and let Speakly transcribe it. Supported formats: .m4a / .mp3 / .mp4 / .aac / .wav / .ogg / .opus / .flac.

Meeting Notes works by capturing system and microphone audio, so it's not tied to any specific meeting platform. Use it with Zoom, Teams, Google Meet, or any other video / voice call, as well as in-person conversations, phone calls, podcast recordings, etc.

Sharing: Once a transcription is complete, you can generate a share link to send to colleagues.

Note: Meeting Notes features consume Genspark Credits (billed by duration) and are not covered by Speakly's free quota.

FAQ

Use Cases

Q: What is Speakly best for?

The following scenarios deliver the best experience:

Email replies — dictate the email and the AI cleans it up into a professional tone; turning on the Professional Rewrite instruction makes the output even more polished and formal
Document writing — think aloud while you speak, generate drafts quickly, reduce the "staring at a blank page" friction
Meeting prep — speak your thoughts aloud and get an outline or talking points
Quick capture — voice memos without switching apps
Multilingual communication — speak in your native language, get text in the target language
Coding assist — use the Terminal Helper instruction to turn natural language into runnable CLI commands

Q: Is it suitable for long, continuous voice input?

Yes. Speakly does not impose a hard cap on single-recording duration — as long as you keep speaking or producing audio, the session stays alive. It's well-suited for writing long documents, recording speech drafts, or post-meeting content cleanup. If there's prolonged silence (around 60 seconds without audio), the session may auto-close.

Q: Can I customize behavior per app?

Yes. Use the Custom Instructions feature to create multiple rule sets and switch between them in different contexts — e.g. use Terminal Helper while coding, Professional Rewrite while writing emails.

Platforms & Compatibility

Q: Which platforms and environments are supported?

Speakly supports both desktop and mobile:

Mac / Windows: full keyboard shortcut experience
iPhone / Android: activated via in-app button; no keyboard shortcuts

Speakly relies on cloud AI for speech recognition and text polishing, so a network connection is required — offline use is not supported.

Q: Does it work in every app on my computer?

Yes. Speakly runs as a system-level tool, so it works in virtually any place that accepts text input — including browsers, the Office suite, Slack, various CRM tools, Zoom chat boxes, Teams message bars, and 100+ other applications.

Pricing & Free Trial

Q: How is Speakly billed? Is there a free quota?

Speakly's features fall into two categories: Dictation-class features are covered by a free quota, while Agent Mode and Meeting Notes consume Genspark Credits.

Dictation-class features (covered by the free quota below):

Default Shortcuts: Dictation and Hands-free Mode
Recommended Shortcuts: all built-in instruction templates (Translate to English, Terminal Helper, etc.)
My Shortcuts: all your custom instructions

Free quota for Dictation-class features depends on user type:

Plus / Pro / Team Plus / Enterprise members: unlimited use, no word or duration limit
First-time users: automatic 7-day free trial, unlimited use during this period
Invite reward: for each new user who successfully uses Speakly via your invite, both the inviter and invitee receive 30 days of free use, stackable on top of existing benefits
Free users after the trial: 4,000 words per week, resets every Monday at 00:00 UTC

Important: the free quotas above (including the 7-day trial and the 30-day invite reward) only cover Dictation-class features. Agent Mode (Ask Genspark) and Meeting Notes are not included — using them consumes Credits even during the trial or invite reward period.

Credit-consuming features (not covered by the free quota):

Agent Mode (the Genspark Agent tasks summoned by double-pressing the shortcut)
Meeting Notes (covers New Note recording, Live Translation real-time transcription + translation, and Import File transcription)

Q: What if I run out of weekly word quota?

When a Free user's weekly quota is depleted, Dictation features will be paused. Options:

Wait for automatic reset next Monday at 00:00 UTC
Upgrade to Plus / Pro for instant unlimited use
Invite a new user to Speakly — once they make their first voice input, both you and the invitee get 30 days of free use

Language Support

Q: Which languages are supported?

Speakly supports 100+ languages, including English, Chinese, Japanese, Korean, Spanish, French, German, Italian, Portuguese, Arabic, Hindi, and many less common languages.

Q: Can I mix multiple languages in one sentence?

Yes. Speakly can handle multiple languages mixed in a single sentence, automatically detecting and transcribing each correctly — no manual language switching required.

Q: Does it work well for non-English languages?

Yes. Speakly delivers solid recognition quality for major non-English languages, and its AI polish step also cleans up non-English content into well-structured text.

Accuracy & Transcription Quality

Q: Speakly is changing my words or adding text I didn't say — what should I do?

By default, Speakly applies AI post-processing to clean up transcriptions — removing filler words ("um," "uh"), fixing run-on sentences, and adding punctuation. Occasionally this can over-correct, especially with idioms, proper nouns, or technical jargon.

If you want certain words preserved exactly as you say them:

Add them to your Dictionary — Speakly will always respect your saved words
Use Send feedback to report the specific issue (see below)

Q: Why does Speakly mis-transcribe names and technical terms?

Proper names, brand names, and specialized terms are the hardest part of any speech recognition system. Speakly's Dictionary feature is designed exactly for this — add names, acronyms, and domain-specific words once, and Speakly will recognize and spell them correctly going forward.

To add a word: left sidebar → Dictionary → Manually Added tab → Add.

Q: Does accuracy improve over time?

Yes. The more words you add to your Dictionary, the better Speakly handles your accent, pace, and vocabulary. Speakly's underlying models also receive regular updates.

Q: How do I give feedback on a specific transcription?

Go to the Home tab → scroll to the bottom Give feedback section → click the Last transcript dropdown to expand the history → select the transcript you want to flag → describe the issue in the input box → click Send feedback.

The feedback includes the transcript and the corresponding audio file to help improve recognition quality. Speakly does not use this data to train general models.

Technical Issues

Q: My shortcut isn't working — what's wrong?

Run through these checks in order:

Confirm the Speakly app is running in the background
Check that the system Accessibility permission is granted to Speakly (Mac: System Settings → Privacy & Security → Accessibility)
Make sure no other app is using the same shortcut
Restart Speakly and try again

Q: The app is unresponsive — what should I do?

Fully quit and relaunch Speakly
Check that your network connection is working
If the issue persists, uninstall and reinstall the latest version

Q: Speakly didn't insert text at my cursor position — what's wrong?

If text isn't going to your cursor, the focus probably shifted away during recording, or system Accessibility permission isn't properly granted.

Steps to check:

Verify Accessibility permission: Mac → System Settings → Privacy & Security → Accessibility → ensure Speakly is checked
Keep the target app focused throughout the recording — don't switch to other windows mid-dictation
If insertion still fails, Speakly will automatically show a popup with the transcribed text so you can copy and paste manually

Q: Where can I see my transcription history?

Open the Home tab to see your full transcription history, grouped by Today / Yesterday / etc. You can search by keyword to find specific transcripts.

Q: How do I disable custom instructions for a specific task?

In Speakly's settings or instruction management page, you can temporarily disable a specific custom instruction or switch to the basic transcription mode (no instruction applied).

Q: Is what I say being saved? Is the data secure?

Speakly transmits and processes voice data over encrypted channels and does not permanently store your raw audio. For full details, see the speakly.ai Privacy Policy.

Q: How do I report issues or send feedback?

You can submit feedback directly inside the Speakly client:

Switch to the Home tab
Scroll to the bottom of the page to find the Give feedback section
Click the Last transcript dropdown to expand the history, then pick the transcript you want to flag (defaults to the most recent one)
Describe the issue or suggestion in the How can this transcript be improved? input box
Click Send feedback to submit

Help Center

SpeaklyTry Now