Your invisible AI, over everything you do.
Ask about anything on your screen — an error, a contract, a chart, an email. Listen to any conversation and get the right answer before you need it. One overlay, no tab, no trace.
What you see. What they see.
The same moment, two screens.
Whatever your day looks like
Pick your job — the same overlay does it. One assistant, not twenty apps.
…and endless more. Pluely adapts to whatever prompt you give it.
Ask about anything on your screen — without switching to a chat tab.
See howThree steps, then it's just there
No setup ritual, no browser tab to babysit.
Install
Download for macOS, Windows, or Linux. A small native app — you're set up in under a minute, no account required to start.
Pick a prompt, attach your files
Curated prompts for common jobs — General, Interview, Coding, Meeting, Translator — or write your own and attach the files that matter: a résumé, a pricing sheet, your docs. Every answer uses them.
Ask, or just listen
Screenshot anything on your screen and ask about it, or start listening and let answers appear on their own as the conversation unfolds. Either way, you never tab away.
Answers arrive before you ask
In any live conversation — a class, a call, a negotiation — Pluely watches the transcript and responds on its own: when it detects a question (even without a question mark), or after every pause. You pick the trigger and the pace.
- Never answers mid-sentence — it waits for the speaker to finish
- Short remarks like “okay, got it” never trigger a response
- Every answer combines your chosen prompt, its attached files, and the live transcript
Hears everything. Joins nothing.
Bot notetakers announce themselves and change how people talk. Pluely captures mic and system audio locally instead — speaker labels, echo cancellation, dozens of languages — and works with every meeting app, lecture stream, or video because it never has to join anything.
Live transcription in detailIt sees what you see
Whatever is on your screen — an error, a chart, a contract, a design — drag a box around it and ask. Attach PDFs and docs up to 20MB. Answers stream in with a Stop button and follow-up chips, so you steer instead of re-prompting.
Everything Ask mode can doReading the screenshot…
200+ models, strongest first
One picker, every frontier model — Claude, GPT, Gemini and more, searchable and sorted strongest-first. Web Search and Perplexity Search sit pinned at the top and return answers with citation chips for questions the training data can't cover.
Browse model selectionPrivate by architecture, not by promise
Bring your own API keys — encrypted on your device — or run through local CLIs like Claude Code, Gemini CLI, OpenAI Codex, Qwen Code, and Ollama. Conversation history stays local. On screen shares the overlay simply isn't there, and it never steals focus from the window you're working in.
Connect your own keys or a CLISee every feature, live
The same overlay, walked through one capability at a time — each with a moment from a real day where it earns its place.
Listen mode captures your microphone and your computer's audio locally and turns them into one speaker-labeled transcript, in real time. Nothing dials into the call, nothing shows in the participant list.
You're four back-to-back calls deep and someone says “thoughts?” — the transcript is already there, so you scroll two lines, see the question, and answer like you never drifted.
Read the full guidePluely watches the live transcript and responds without being asked: when it detects a question, or after every pause — your choice, with pace control. Detection runs on your device and is free; you only spend an AI request when a response is generated.
In an interview, “walk me through your last project” lands and a suggested answer is already streaming before your pulse settles — built from your selected prompt and your actual résumé.
Read the full guideThe buttons above the Listen composer are your own system prompts — up to six, re-orderable, fully editable. The button shows two words; tapping it runs the full prompt, plus its attached files, against the live conversation.
The call is wrapping and someone asks for next steps — you tap “Quick Recap” and get decisions, owners, and blockers as a clean list you can read straight into the mic.
Read the full guideAsk mode answers in place, over whatever app you're in. Attach a full screenshot or drag-select a region — an error, a chart, a form — and the answer streams into the overlay without you switching tabs.
A build fails with a stack trace you've never seen. One capture, “what's breaking here?”, and the fix — with the exact command to run — arrives before you'd have finished writing the search query.
Read the full guideAttach files from your library to any system prompt: résumé, product spec, pricing sheet, course notes. Whenever that prompt is active — in Ask, in Listen, in automatic responses — the AI answers with those documents in hand.
A prospect pushes back on price mid-call. Your sales prompt carries pricing.pdf, so the suggested reply quotes the right plan, the right number, and the annual discount — no alt-tabbing to find the sheet.
Read the full guideToggle “Use image” and each question you send grabs a fresh screenshot first — the AI always sees what you see, with nothing to attach. Auto-screenshot can also fire once or on an interval.
An email needs a careful reply. With Use image on, you type “reply to this for me” — no copy-paste, no context-setting — and a draft appears that references exactly what's on screen.
Read the full guide↑ The actual overlay, playing each feature — it advances on its own, or click a feature to jump.
Why Pluely?
Plenty of tools summarize your day. Only one sits invisibly over it and answers in the moment — on your models, your keys, your machine.
Live, not after the fact
Most tools summarize the conversation once it's over. Pluely works during it — the answer is on your screen while the question still hangs in the air.
Automatic responsesInvisible by design
No bot in the participant list, nothing on screen shares or recordings, and it never steals focus. Discretion isn't a setting — it's the architecture.
Stealth & privacyOne overlay for everything
Ask about anything on your screen; listen to any conversation in any app. One assistant that follows your whole day, instead of a different tool per task.
Ask modeYour AI, your way
Run on 100+ managed models with zero setup — or bring your own API keys and local CLIs like Claude Code, Gemini CLI, OpenAI Codex, Qwen Code, and Ollama. You're never locked to one provider.
Connect your ownLocal-first privacy
Conversations, transcripts, and files live on your device. Keys are encrypted on-device. Only what's needed to answer a request you make ever leaves your machine.
Local-first privacyFree forever, private forever
The free plan isn't a trial, and your data — chats, meetings, files — is stored and encrypted locally. It never leaves your computer, so every claim on this page holds up on your own machine.
Compare alternativesEvery feature, explained
The complete set — each one links to its docs page, so nothing here is a mystery.
Automatic responses
Fires on detected questions (no “?” needed) or after every pause. You pick the trigger and pace — Fast, Balanced, Relaxed — and it never answers mid-sentence.
Learn moreLive transcription
Mic + system audio captured locally with Speaker 1/2 labels and echo cancellation. Works with every meeting app because no bot ever joins.
Learn moreTranscription languages
Dozens of languages with multilingual auto-detect — transcribe calls, lectures, videos, or podcasts in whatever language they arrive in.
Learn morePrompts + knowledge files
Curated defaults (each paired with a recommended model) plus your own custom prompts. Attach a résumé or pricing sheet and every answer is grounded in it.
Learn moreAI prompt generator
Describe the task in a sentence and get a complete custom system prompt back — no prompt-engineering homework.
Learn moreAsk mode
Streaming answers with a Stop button, follow-up chips, and Latest/All conversation views — over whatever you're doing.
Learn moreScreenshots & vision
Full capture or drag a box around just the part that matters — an error, a chart, a paragraph. “Use image” attaches your screen to every message.
Learn moreInterval screen monitor
Every 15s to 5m, Pluely screenshots your screen and answers automatically — an ambient “watch this and flag anything important” copilot.
Learn moreDocument Q&A
Attach PDF, DOC/DOCX, PPT/PPTX, TXT, MD, CSV, or images up to 20MB and ask direct questions. Extracted text persists across restarts.
Learn moreFiles library
Upload once, reuse anywhere. Processed documents are saved with their extracted text — re-attach without spending another upload.
Learn morePush to talk
Hold ⌃Space, speak your question, release. Dictation beats typing when your hands are mid-task.
Learn moreWeb search with citations
Two pinned entries — Web Search and Perplexity Search — answer from the live web, with citation chips linking each fact to its source.
Learn more100+ models
A searchable picker, strongest models first — Claude, GPT, Gemini and more. Switch per question with one tap.
Learn moreOne-tap transforms
Make concise, Make longer, Explain, Summarize, Key points (decisions + action items), Translate — on any answer, transcript, or document.
Learn moreResponse settings
Set response length (concise ↔ detailed) and response language once — every answer, app-wide, follows your preference.
Learn moreNotes
Append AI summaries into an editable notes doc as you go — the meeting ends with the write-up already started.
Learn moreBring your own keys
Add your own provider API keys — encrypted on your device, unmetered by Pluely. The free plan runs entirely this way.
Learn moreLocal CLIs
Route through Claude Code, Gemini CLI, OpenAI Codex, Qwen Code, or Ollama running on your machine — your existing subscriptions and local models, in the overlay.
Learn moreMeetings
Every session is saved and reopenable, with one-tap summaries, key points, and exports when the call ends.
Learn moreHistory & library
Local-first conversation history on your device — search it, reopen it, and nothing needs a cloud account.
Learn moreStealth
Invisible on screen shares and recordings, never steals focus, and app-icon stealth keeps it out of sight entirely.
Learn moreOverlay control
Draggable, always-on-top, resizable, with adjustable transparency and reorderable Ask/Listen tabs — it sits exactly where you want it.
Learn moreGlobal shortcuts
Summon, hide, capture, and ask from anywhere — every shortcut is customizable.
Learn moreTransparent usage meters
Live countdowns and monthly budgets visible in the app — you always know exactly what's left before you hit a limit.
Learn moreThe lightweight alternative to Cluely
A 9–16 MB native app with a free plan forever — not a heavyweight subscription you have to justify before you've even tried it.
Start free. Upgrade when it earns it.
The free plan is the full product with your own keys — paid plans add bundled models and transcription.
Bring your own API keys. Full overlay, full stealth, no account.
- Invisible in calls and screen shares
- Your own AI + speech providers
- Custom prompts and profiles
- Zero data collection
Bundled AI and speech-to-text — nothing to configure.
- 251 Pluely AI models built in
- 5 hours live transcription / month
- 300 AI requests / month
- CLI included — Claude Code, Gemini CLI, OpenAI Codex, Qwen Code, Ollama
Everything in Pro, with much bigger monthly budgets.
- 30 hours live transcription / month
- 2000 AI requests / month
- 30 document uploads / month
- 2 licenses (2 device activations) + priority support
Bringing your own providers? Dev Pro — $120 one-time unlocks every Pro feature for life for BYO-keys users — no subscription, CLI included, and every new release before it goes public.
The questions everyone asks first
Put an answer over everything.
Install it now — it'll be there for the next error, the next call, the next thing you'd rather not figure out alone.
Free plan forever · Invisible on screen shares · No bot joins your calls