SuperAPI AI Directory

Ethertext

AI-powered clipboard for text transformation and productivity.

Open website

About this tool

Ethertext is an AI clipboard tool (Your AI Clipboard), and its core workflow is "Copy. Transform. Paste.", using AI-powered text editing to improve everyday work efficiency. It is available for macOS and handles various types of text processing through keyboard shortcuts and clipboard operations.

Purpose and Positioning Ethertext embeds AI text processing into the clipboard workflow, allowing users to rewrite text, process code, memorize information, and recall it without switching to complex applications. It is officially described as "One-Click Wonders": with a single click, text can go from "good" to "better."

Main Capabilities

  1. Text transformation: Select any transformation option from the dropdown menu to apply AI processing to the text in the clipboard. It is officially said to allow you to choose the tone and style that fits your needs, making the text suit your personal requirements.
  2. Code processing: Explain, debug, or translate code snippets with "laser-like precision."
  3. Memorize and Recall: Ethertext can memorize text for later recall (Memorize and Recall); update notes mention that Recall can use embedding models to improve results.
  4. Voice and screen: It can record, transcribe, and memorize voice; it can take screenshots and memorize text on the screen. The update log mentions the use of Apple's local Vision model to transcribe screen text, as well as Apple's local Audio model.
  5. Multi-model support: Supports API keys for OpenAI, Google Gemini, and Anthropic; it also supports Ollama local AI and MLX on-device models. The update log shows that previously supported models included GPT-5.2, Gemini 3, Claude 4.5, GPT 4o mini, and others.

How to Use Basic steps:

  1. Use ⌘C to copy the source text to the clipboard.
  2. Click Transform and select a transformation from Ethertext's dropdown menu.
  3. After hearing the prompt sound, use ⌘V to paste the text.

Common shortcuts:

  • ⌘⇧T: Clean up text; or fix the selected text with one click.
  • ⌘⇧M: Memorize text or a webpage.
  • ⌘⇧D: Record, transcribe, and memorize voice.
  • ⌘⇧S: Take a screenshot and memorize the text on the screen.
  • ⌘⇧R: Recall past text and ask questions about it.

API key configuration: Add keys in Settings, including:

  • OpenAI:https://platform.openai.com/account/api-keys
  • Google Gemini:https://makersuite.google.com/app/apikey
  • Anthropic:https://console.anthropic.com/settings/keys

Local AI (Ollama): Install Ollama and download a model, then enable Ollama support in Settings and select a model. Official example commands: brew install ollama; ollama pull llama2.

Installation Methods

  • Install via Homebrew: brew install thrashr888/ethertext/ethertext
  • It can also be downloaded directly. The official update log mentions that version 1.13 added Homebrew cask support, and the release process switched to DMG installers.

Use Cases Ethertext is suitable for quickly rewriting selected or clipboard text in any application, adjusting tone and style, explaining and debugging code, translating code, memorizing and recalling text, transcribing voice, extracting screen text, and more. It is a good fit for macOS users who want to use AI to process text without leaving their current workflow. Local model support (Ollama / MLX) also provides an option for users who value privacy or want to run offline.

Limitations and Information Not Yet Provided

  • In terms of platform, the materials mainly describe the macOS experience; support for other platforms is not yet provided.
  • Pricing, cost, and subscription plans are not yet provided.
  • Free quotas and usage limits are not yet provided.
  • Specific comparisons of the effects of different models are not yet provided.
  • Apart from macOS 14+, compatibility with other system versions is not yet provided.
  • If using cloud models, you must provide the API key for the corresponding service yourself; local models must be installed and downloaded yourself, and specific hardware requirements are not yet provided.