Keyboard Shortcut
Ctrl+Shift+D starts and stops a recording, even when Saytrue is running in the background.
Speak or type what you want to say.Saytrue suggests a new version and tells you in a few words what it does differently.It draws on approaches from psychology, such as the Four-Sides Model.
Each lens looks at your text from a different angle.
Shows which feeling comes through in your text.
Shows how your text sounds: how formal, how close, how direct, how lively, and what mood it carries.
Finds flaws in reasoning, such as hasty generalizations or attacks on the person.
Shows what you observe, what you feel, what you need and what you are asking for.
Finds thought patterns such as black-and-white thinking, catastrophizing or overgeneralizing.
Shows what your sentence says about the matter, about you and about your relationship, and what it asks for.
Shows whether you are lecturing, being defiant or speaking as an equal.
Assigns your text to an area of life, such as work, health or relationships.
Shows how you see the situation: how new, how important, how manageable and how well it fits your values.
Shows whether you want to achieve something or prevent something.
What will I do differently next time?
The right wording, without having to know the models behind it
Ctrl+Shift+D starts and stops a recording, even when Saytrue is running in the background.
Three versions of the same statement: direct, empathic and de-escalating. Also for text from other programs, via Ctrl+Shift+R or, on macOS, the Services menu. Saytrue takes the form of address from the text; otherwise you choose informal or formal.
One version per selected lens, each from a different psychological perspective.
“Observation, not judgment”, “Adult-to-adult”: you see what the version does differently, without model names.
Browse, listen to and export earlier inputs.
Answers questions about the app and its approaches. The knowledge it draws on is built into the app.
Type, speak into the microphone or load an audio file.
Ctrl+Shift+Q records and puts only the text on the clipboard, without analysis and without saving. On macOS and Linux.
Three steps to more precise words
Speak, type or load an audio file.By default, Saytrue turns speech into text on your device.
A lens is a psychological perspective, such as Nonviolent Communication or the Four-Sides Model.Set each lens to Off, Analyze or Suggest.
With Suggest, Saytrue rephrases your text.You decide what to use.
You decide where your texts are processed and where they go.
History and recordings as PDF, Word or Markdown; recordings also as JSON
On your device with Ollama or Apple Intelligence, in the cloud with OpenAI, Anthropic or Mistral (EU)
Access keys for cloud providers are kept in the operating system's credential store (Keychain, Credential Manager, Secret Service, Android Keystore)
Without a cloud provider, all data stays on your device
Free makes your tone warmer.Pro uses all lenses and Coaching.
Everything in Free, plus:
System requirements, installation, and setup
Which analysis runs with which local model in which language, as shown in the app.
View Overview →sudo snap install saytrue Snap is available for x64. For ARM64, use the .deb, .rpm or AppImage package from GitHub.
Saytrue requires Ollama, Apple Intelligence (macOS 26+), or a cloud provider (OpenAI, Anthropic, Mistral). For optional transcription, the app downloads a speech model for whisper.cpp on first launch.
Setup Guide →