Backlog

Next up

In Progress

Done

Main Roadmap

Support showing app interface text in the supported languages

In Review

App Interface: Language Support

Whisperstream

Allow a dictionary to be imported/exported via a CSV file.

Dictionary import/export

Bass

Dark Mode

Linux Support

macOS Support

Support additional languages either using Qwen ASR or Whisper V3 Turbo

Planned

Transcription: Additional Language Support

Allowing keypress simulation instead of clipboard input to support Citrix and RDP environments.

Citrix + RDP Support

For users without a GPU, the local cleanup option is too slow. The only option is currently cloud with a free API key. We’re working on training a model that allows local cleanup for everyone with a small enough model to run on a CPU

Fully Local Cleanup (CPU users)

Use your own API Key from a cloud provider or Ollama for local and use an LLM to post process your output and format it in the way that you can.Supports app detection to auto activate modes.

Completed

LLM Post Processing aka. Enhancement

Allow LLM post processing with a built in model for users with a GPU. Chat and Code prompts are already tuned to work well with this model.

Fully Local Cleanup Enhancement (w/ GPU)

Allow detecting which app is currently in focus and choosing a post processing prompt accordingly.

App Detection & Modes

Searchable local history of past dictations. Store past transcripts, add a screen that shows them all, allow naming them and searching by title or contents.

whisperstream

Main Roadmap