Landmark DocsDashboard →

On stream

Text to speech

Landmark reads messages out loud on your stream. Three switches decide what gets read. Each has a voice. Pick where the sound comes out. That is the setup.

Reads aloud

Chat messages
Every chat message, on every platform you connect, in the voice you pick. You can limit it to followers, subs, or mods.
Read my message
A redeem. Viewers redeem it with coins on your Landmark page, or with bits, channel points, a gift or a direct tip on the platforms you set, and their words are read in the voice you pick for it. Prices and platform rules live on the card, like every other redeem.
Subs, gifts and raids
Says a line for each, like {user} just subscribed!. You can change the lines.

Turn a switch on, pick its voice, done. One master switch at the top turns everything off without losing your settings.

Plays in

My Control Dock: open the dock and it reads aloud. Nothing to add to OBS. Or My OBS overlay: add the Speech element to your overlay and the reads play there. Reads wait in one line and play one after another. Chat only joins the line while the dock or overlay is open, so nothing piles up while you are away.

Filters

Four word groups you can turn on: slurs, sexual, swears, violence. Slurs are on for everyone. Type a message into the Try it box and it shows you exactly what would be read, or why it would be blocked. Under More filters: swap one word for another, remove links, calm down shouting, block a viewer, and the pacing between reads.

Viewers

Viewers can pick their own voice by typing !voice and a name in chat, if you let them. They can only pick free voices, so a viewer can never run up your bill.

What it costs

Two kinds of voices. The difference is cost, and the page tells you which is which.

Free voices
Brian and about a dozen classic voices. Free up to 1,000,000 characters a month, then $0.0044 per 1,000 characters.
AI voices
Lifelike voices. $0.0125 per 1,000 characters, every read. An AI voice is marked AI wherever you pick one.
Novelty voices
Animalese, Sans, Blip, Morse. Made in your browser. Never bill.

The Usage pill at the top of the settings page shows what this billing period cost so far. Only what is actually read bills. A blocked or skipped message bills nothing. The full rates are on the fees page.

Hear the voices

Each clip says the same line so you can compare. Press play.

Animalese
0:00 / 0:00
welcome to the stream
Chirpy, one chime per letter. Free, made in the browser.
Sans
0:00 / 0:00
welcome to the stream
One deep blip per syllable. Free, made in the browser.
Tomodachi
0:00 / 0:00
welcome to the stream
A Mii-style sing-song voice. Gold only. Bills per read.
DECtalk
0:00 / 0:00
this is the dectalk robot voice
The classic 1980s robot voice. Bills per read.

Every voice is on the Voices page with a play button.

For the curious

Advanced rules

Under the settings there is an Advanced rules list. Every read follows it, top to bottom, and the first rule that matches wins. The three switches set these rules for you. Open the list only when you want something specific: a paid read with a price per message, a rule for one platform, a rule for one person, or a different voice for cheers than for tips. Reset to the switches rebuilds the list from the switches if it ever gets tangled.

Voice commands in chat

Tomodachi and Animalese read a short command at the start of a message and change how they sound. Put it in square brackets, then a space, then the message. The command is not read out.

High and fast
[p85,s70] hello chat, i am wide awake
Low and slow
[p15,s30] greetings... from the deep
Tomodachi preset
[v3] try voice three through [v8], and [v5,p80] for a preset plus a tweak

p is pitch and s is speed, 0 to 100. Tomodachi also reads q quality, t tone, a accent and i intonation.

[p85,s60] welcome to the stream
0:00 / 0:00
welcome to the stream
High pitch, a bit faster.
[p20,t30] welcome to the stream
0:00 / 0:00
welcome to the stream
Low pitch, dark tone.

Make Tomodachi sing

Start the message with [sing], then write each word with the note it lands on. A note is a letter A to G, an optional #, and an octave, like C4. Add :2 to hold a note.

[sing bpm150] ma{E4} ry{D4} had{C4} a{D4} lit{E4} tle{E4} lamb{E4:2}
Tomodachi singing
0:00 / 0:00
The notes above: "Mary had a little lamb". bpm sets the speed.

Elements can speak

Any element you build can call landmark.tts(text) to say something out loud, like when a goal is hit. It bills like any other read. See the element API.

Open Text to speech →

Last updated · June 28, 2026