Trakktor

Trakktor is an experimental Rust utility for running AI tasks. The tasks are written in JavaScript and executed by a single binary.

Instead of relying on Node.js, Trakktor embeds the Boa JavaScript engine and provides a small host API (HTTP, HTML→Markdown, AI chat, file IO, preview, Text-to-Speech) to your JS scripts.

The goal is a single, self-contained executable that can run AI automations without external language runtimes.

Status

Early development.

It works, but please don’t use it for anything important yet. The main purpose right now is to explore the approach and iterate on the runtime API.

Key ideas

• Rust binary, JS tasks: tasks are regular .js files.

• Embedded JS engine: runs on Boa (no Node.js).

• Interactive previews in the terminal (TUI): tasks can emit artifacts via preview, and Trakktor can render them in an interactive terminal UI (currently audio), powered by the fantastic Ratatui.

  

• Minimal (for now) host API exposed to JS:

  • httpGetText
  • htmlToMarkdown
  • chat
  • readFile
  • preview
  • textToSpeech

  (6 functions today; more will be added as the project evolves).

One executable, no language runtimes

Trakktor is designed to run as a single Rust binary:

• No Node.js / Python (no external language runtimes).
• JS execution is embedded via Boa.

In other words: you run one executable, and it runs your JS task.

Quickstart

To see what exists today, you need Rust, then:

git clone https://github.com/lymar/trakktor
cd trakktor

export OPENAI_API_KEY=...

cargo run --bin dev_run_js -- examples/text_to_speech.js

Included examples

• examples/on_this_day.js: fetches Wikipedia, converts HTML→Markdown, asks the model to extract the “On this day” section into JSON.
• examples/structured_output.js: minimal example of requesting JSON output.
• examples/text_to_speech.js: generates audio via textToSpeech, then opens an interactive TUI preview.
• examples/audio_preview.js: reads a local audio.mp3 via readFile, then opens an interactive TUI preview.

License

MIT