Your unfair advantage on every technical call
Let Me Solo IT listens to your call and shows you a short hint instantly, written the way you would say it, with the key points and copy-ready code. In a small private window over your call.
Windows, macOS and Linux. A ready-made build for each.
The silence after a hard question is where a call is won or lost. You are the one who never loses the thread: the answer, the points they care about and the code, the moment they ask.
Real screenshots from the app, on one sample call: a plan renewal in which the customer’s engineer also asks about backend architecture.
The app transcribes your microphone and the call’s audio. When a question lands, a short hint appears instantly, written the way you would say it out loud, and fills in as the AI writes, with the one to three things the other participants are really asking about. A few self-hosted AI endpoints answer all at once instead. The gauge in the header reads how the conversation is going.
Ask for a snippet by voice or by typing. Code in a hint lands in its own box, labeled with its language, with and text you can select. Drop a diagram or a screenshot into the input and the assistant answers what is in it.
On Windows and macOS the app’s window is excluded from your own screen shares and recordings, so your notes stay yours. On macOS a sharing or recording app can override this when it captures the whole display without window filtering (in Zoom’s settings, under Share Screen → Advanced → Screen capture mode, the mode “Advanced capture without window filtering”); the default modes respect it. On Linux the window is visible like any other window. It never hides the assistant from the other participants: you tell everyone on the call, and the app asks you to confirm that before the first call.
One press re-asks the last minute of the conversation as a single question, which fixes a question that arrived in fragments. and jump between the participants’ questions, and the feed waits while you read back.
Every keyboard shortcut can be changed. Keep a saved profile for each kind of call, with a few lines about you and your role, so hints draw on your real experience, and switch them mid-call. Use your own AI account, through an API key or a command-line tool you are already signed in to, a local model, or the AI included in a hosted plan.
The app transcribes the other participants’ voices, and in many countries and several US states recording or transcribing a conversation needs the consent of everyone in it. Tell the room before you start, and get their consent where the law requires it. Before the first call the app shows its disclaimer and asks you to confirm exactly that. Nothing is recorded until you accept.
Click a box to read more.
One press captures the screen you choose and asks for help with it and the last minute of the call.
Queue a picture of the screen you choose, up to 10 of them, and send them together with a typed question.
Attach a chart, a slide, a code task or a document (.txt, .md, .pdf, .docx, or .doc where LibreOffice is installed, up to 20 MB each, up to 10 per question) and ask about it. Documents are read on your computer.
Jump between the participants’ questions; the feed resumes at the latest one.
English is the default: speech is read in exactly the language you pick, with no detection. Automatic detects the language of each side of the call with speech on your computer or our hosted speech and answers from our hosted model, and the hints follow yours. The twelve: English, Ukrainian, Russian, German, Spanish, French, Japanese, Polish, Swedish, Chinese, Korean and Portuguese; Ukrainian, Chinese, Korean and Portuguese need a graphics card or Apple Silicon to be heard on your computer.
With a language you pick, one press switches speech and hints to a second language, and the next press switches back.
One press lists what was asked, what was answered or agreed, and the next steps. Clear chat starts a new chat, and can summarise the one it clears first (off by default).
A new question is answered at once, beside the hints still being written, up to three at a time.
The app stops hearing you while it still hears the other participants, and drops a phrase under way. It does not mute you in the call, and every run starts unmuted.
The app stops hearing the other participants while it still hears you: nothing of theirs is transcribed, sent or charged. Every run starts with them heard.
Each chat is backed up on your computer as it happens and never uploaded, for 30 days by default.
Save the chat to your Documents folder as TOML, Markdown, HTML, PDF, plain text or a folder; it is saved on its own when it ends, unless you turn that off. The HTML page opens offline in any browser, with a list of the turns to jump to.
Load an export or a backup back into the window, and ask follow-up questions, rewind or summarise it as if the call were still on.
Pick up to ten earlier calls, with the same company for instance: what was said in them goes with every hint, so the hints build on it.
Pick the microphone and the speaker the app hears; the change applies at once.
Let clicks pass through the window to the app underneath while the hints stay visible.
Type a question in the window without saying a word, and send it from anywhere with a shortcut.
Ctrl+Alt with S for Help me, P to capture a screen, A to send now, R to rewind, [ and ] between questions, U for a summary, N for a new chat, M to mute, O to pause the participants, K for the second language, E to export, X for click-through, H to hide the window and Q to quit. Add Shift and an arrow key to move the window. Change any of them in the settings.
Use your own API key, a signed-in command-line tool or a local model, or take credits for the hosted model and hosted speech-to-text.
Use your own API key or a local model, for good.
Use your own API key or a local model.
Hosted answers and speech, with credits every month.
Credits for hosted answers and speech, as many packs as you need.
| Own Key Lifetime | Own Key Monthly | All-Inclusive | Credit Pack | |
|---|---|---|---|---|
| Use your own AI: an API key, a signed-in command-line tool or a local model | Included | Included | Included | Not included |
| Credits for the hosted model, no account or key needed | Not included | Not included | 10,000 every month | 1,000 per pack, for 180 days |
| Hosted speech-to-text, paid in credits | With a Credit Pack | With a Credit Pack | Included | Included |
| Works without internet | Up to 30 days at a time, with a local model and local speech | Not included | Not included | Not included |
| How long it lasts | For as long as we offer the app | Renews every month | Renews every month | Until its credits are used or expire |
| Cancel to stop renewing | Bought once: nothing to cancel | Included | Included | Bought once: nothing to cancel |
All sales are final; credits are non-refundable. Own Key Monthly, Own Key Lifetime and All-Inclusive work on 2 computers; a Credit Pack alone works on any number. The app shows how many credits are left.
Pick your system: each has its own build and steps.
Windows requirements: Windows 10 21H2 or Windows 11, 64-bit (x64).
Get the Windows build, LetMeSoloIT-1.0.1-setup.exe.
Run the installer and follow its steps. The build is unsigned, so on the first run SmartScreen shows "Windows protected your PC": click More info, then Run anyway.
Open Let Me Solo IT from the Start menu and allow the microphone when asked; the switch lives in Windows Settings, then Privacy & security, then Microphone, for desktop apps. Before the first call, the app asks you to confirm that you will tell everyone on the call.
The PowerShell script from this site downloads the build for your system, checks it and installs it for your user only, without administrator rights: a silent setup into %LOCALAPPDATA%\Programs\Let Me Solo IT.
Install, or update to the latest release (add -Force to install it again):
irm https://letmesoloit.com/install.ps1 | iexUninstall; your settings, sign-in, logs and speech models stay:
& ([scriptblock]::Create((irm https://letmesoloit.com/install.ps1))) -UninstallmacOS requirements: macOS 13.3 Ventura or later on Apple Silicon.
Get the macOS build, LetMeSoloIT-1.0.1.dmg.
Open the .dmg and drag LetMeSoloIT.app into Applications; if an older copy is there, choose Replace. Keep one copy only: with two, macOS refuses to open the app ("not responding"). The app carries only an ad hoc signature, so macOS refuses the first open: go to System Settings, then Privacy & Security, find the message naming LetMeSoloIT near the bottom, choose Open Anyway and confirm (on macOS 13 and 14, right-click the app, Open, then Open works too). That is needed once, not every launch. After every update, allow Screen Recording and Microphone again (remove the app from each list with the minus button if the switch looks on, and add it back to Screen Recording with the + button from Applications), then quit and reopen the app.
Open the app from Applications and grant the permissions it asks for, Microphone and Screen Recording. macOS applies Screen Recording after a restart, so quit and reopen the app once. Before the first call, the app asks you to confirm that you will tell everyone on the call. No sound from the other participants? See the macOS help.
The shell script from this site downloads the build for your system, checks it and installs it for your user only, without sudo: the app in ~/Applications, or it updates the copy already in /Applications.
Install, or update to the latest release (add --force to install it again):
curl -LsSf https://letmesoloit.com/install.sh | shUninstall; your settings, sign-in, logs and speech models stay:
curl -LsSf https://letmesoloit.com/install.sh | sh -s -- --uninstallLinux requirements: A 64-bit Linux distribution with glibc 2.34+.
Get the Linux build, LetMeSoloIT-1.0.1-x86_64.AppImage.
Make the AppImage executable. It carries everything it needs, so there is nothing else to install.
chmod +x LetMeSoloIT-1.0.1-x86_64.AppImageOnly if it stops with "Cannot mount AppImage, please check your FUSE setup." (a minimal system or a container), install FUSE 3:
| Distribution | Install |
|---|---|
| Debian, Ubuntu, Linux Mint, Pop!_OS | sudo apt install fuse3 |
| Fedora, RHEL, Rocky, AlmaLinux | sudo dnf install fuse3 |
| Arch, Manjaro, EndeavourOS | sudo pacman -S fuse3 |
| openSUSE | sudo zypper install fuse3 |
Or, in the same case and without installing anything, start it unpacked to a temporary folder; it starts noticeably slower:
APPIMAGE_EXTRACT_AND_RUN=1 ./LetMeSoloIT-1.0.1-x86_64.AppImageRun it to open the app window. On Linux the window is a normal window, not excluded from screen capture. On GNOME the app asks once which screens it may capture for screenshots: choose all of them. Before the first call, the app asks you to confirm that you will tell everyone on the call.
./LetMeSoloIT-1.0.1-x86_64.AppImageThe shell script from this site downloads the build for your system, checks it and installs it for your user only, without sudo: the AppImage in ~/.local/bin, with its menu entry and icon.
Install, or update to the latest release (add --force to install it again):
curl -LsSf https://letmesoloit.com/install.sh | shUninstall; your settings, sign-in, logs and speech models stay:
curl -LsSf https://letmesoloit.com/install.sh | sh -s -- --uninstallEverything you need to know about the product and billing. The guide covers every feature step by step.
Windows 10 21H2 or Windows 11 (x64), macOS 13.3 Ventura or later on Apple Silicon, or 64-bit Linux with glibc 2.34 or newer (Ubuntu 22.04, Debian 12, Fedora 35 or later). The app window works on all three; on Linux it is not excluded from screen capture. More in the guide: System requirements.
These are estimates. We measured two machines (a 2019 entry-level AMD laptop and a 2019 AMD laptop with a GeForce card) and extrapolated to similar hardware by graphics compute and memory. Intel graphics and Apple Macs are not measured yet. On a plan that includes hosted speech, Automatic uses it first, so the local rows matter only when the local model runs: without hosted speech, or when it stands in for it.
| AMD | Intel | NVIDIA | Apple | |
|---|---|---|---|---|
| Hosted speech (your computer does no recognition) | Ryzen 3 2000 series or newer | Core i3 7th gen or newer | — (no graphics card needed) | Any Apple silicon Mac (M1 or newer) |
| Local speech, Fast model (text about 3.5–5 s after a phrase ends; a phrase takes longer while the language is detected) | Ryzen with Radeon Vega 3 graphics or newer integrated graphics (Vega 8: about 2 s per phrase) | Core 11th gen or newer with Iris Xe graphics (not measured) | Any card from the next row | M1 or newer (not measured) |
| Local speech, Balanced model (about 1.3 s per phrase, about 2.1 s while the language is detected) | Radeon RX 5500 / RX 6500 XT or newer, 4 GB or more | Arc A380 or newer (not measured) | GeForce GTX 1650 / RTX 3050 or newer, 4 GB or more; the model stays on the card the whole call | M1 Pro / M2 Pro or newer (not measured) |
With our hosted speech (re-measured on 2026-10-14, the call played aloud): on the Ryzen 3 3200U laptop the app used at most about 330 MB of memory and 20–27 % of one processor core, and the first words showed about 2 s after someone started speaking; on the Ryzen 5 3550H PC, at most about 300 MB and 23–33 % of one core, and about 2.4 s. Right after the language changes, the first words can take up to about 7 s.
All tiers: a 64-bit processor with at least 2 cores and 4 threads, and 8 GB of memory (the app itself uses up to about 480 MB, and on built-in graphics the Fast model takes about 620 MB more of the shared memory; the rest is for your call app and the system). If local speech is too slow, the app says so. On a plan that includes hosted speech it suggests Fast or hosted speech, or hosted speech alone when Fast is too slow; without it, Fast or the "Use the GPU" switch where it is offered, or the switch alone when Fast is too slow.