A Chrome extension that turns rough ideas into well-structured, effective prompts for LLMs — with a bonus automatic RTL/LTR text-direction fix for Persian and Arabic text on AI chat platforms.
Prompt Enhancer adds an Enhance button next to text inputs on your favorite AI chat platforms. When clicked, the selected provider (OpenRouter or Groq) refines the text into a strong, self-contained prompt and inserts it back into the input — ready to send. It also detects Persian/Arabic text on the page and fixes its direction and alignment automatically.
- ✨ Prompt Enhancement — restructures rough input into clear, actionable prompts while preserving intent, language, and tone
- 🎛️ Enhance styles — 6 modes with tuned prompts and temperatures: Concise (default), Standard, Detailed, Coding, Writing, Reasoning
- 📝 Custom instructions — optional preferences (up to 1000 chars) applied on top, e.g. "Always ask for a table"
- 🔀 Provider Choice — 5 free-tier providers (Groq, OpenRouter, Gemini, Cerebras, Mistral) in one dropdown, each with its own key, model, and free-key signup link
- 🧠 Model Picker — custom dropdown with badges (Default / Free / Fast / Versatile / Reasoning) for each provider
- 💾 Secure Storage — API keys and preferences saved per provider via
chrome.storage.sync - 🌐 RTL/LTR Auto-Fix — Persian/Arabic text is always detected and rendered right-to-left with the bundled Vazirmatn font (CSP-safe, no network fetch)
- ⏸️ Enable / pause switch — header toggle in the popup instantly shows or removes all in-page buttons
- ↩️ Multi-level Revert + Copy — up to 10-step undo history per input, plus a one-click Copy button
- 🎯 Selection-aware — uses
selectionStart/Endfor textareas and containment-checked ranges for rich editors, so only the selected text inside the input is enhanced - ⌨️ Keyboard Shortcut —
Ctrl+Shift+E(orCmd+Shift+Eon macOS) enhances the focused input, with fallback to the first visible input - 💬 Inline Feedback — gradient button with a spinner while processing, plus glass-style success/error toasts
- 🛡️ Robust editors support — works with React textareas and ProseMirror / Lexical / Draft-style rich editors; 60s request timeout with friendly errors (invalid key, credits, rate limit, unknown model, network)
- 🧹 Clean output — strips code fences, wrapping quotes, and "Here's your prompt" preambles
- 🎨 Modern UI — dark-first popup with gradient accents and full light-mode support
Popup — provider, model, and API key settings:
Enhance Prompt button next to the input:
| Model | Model ID |
|---|---|
| Qwen 3.8 27B (default) | qwen/qwen3.8-27b |
| Llama 3.3 70B (Versatile) | llama-3.3-70b-versatile |
| Llama 3.1 8B (Fast) | llama-3.1-8b-instant |
| Llama 4 Maverick 17B | meta-llama/llama-4-maverick-17b-128e-instruct |
| Llama 4 Scout 17B | meta-llama/llama-4-scout-17b-16e-instruct |
| Kimi K2 (Reasoning) | moonshotai/kimi-k2-instruct |
| GPT OSS 120B | openai/gpt-oss-120b |
| GPT OSS 20B (Fast) | openai/gpt-oss-20b |
| Qwen 3 32B | qwen/qwen3-32b |
| Qwen 3.6 27B | qwen/qwen3.6-27b |
| Model | Model ID |
|---|---|
| OpenRouter Free Router (default) | openrouter/free |
| Dots3 Note Preview | dots-studio/dots-3-note-preview:free |
| Liquid LFM 2.5 2.6B | liquid/lfm-2.5-2.6b:free |
| NVIDIA Nemotron 3.5 Lightning | nvidia/nemotron-3.5-lightning:free |
| Thinking Machines Inkling | thinkingmachines/inkling:free |
| Poolside Laguna S 2.1 | poolside/laguna-s-2.1:free |
| Cohere North Mini Code | cohere/north-mini-code:free |
| GLM 5.2 | z-ai/glm-5.2:free |
| MiniMax M3 | minimax/minimax-m3:free |
Gemini (free tier — key at aistudio.google.com)
| Model | Model ID |
|---|---|
| Gemini 3.8 Flash (default) | gemini-3.8-flash |
| Gemini 2.5 Flash | gemini-2.5-flash |
| Gemini 2.5 Pro | gemini-2.5-pro |
Cerebras (1M tokens/day free — key at cloud.cerebras.ai)
| Model | Model ID |
|---|---|
| GPT OSS 120B (default) | gpt-oss-120b |
| Llama 3.1 8B | llama3.1-8b |
Mistral (free plan — key at console.mistral.ai)
| Model | Model ID |
|---|---|
| Mistral Small (default) | mistral-small-latest |
| Mistral Large | mistral-large-latest |
| Codestral (Code) | codestral-latest |
| Mistral Nemo | open-mistral-nemo |
💡 Saved models that are later removed by a provider automatically fall back to the default — no stale-model errors.
ChatGPT (chatgpt.com, chat.openai.com), Claude, Gemini, AI Studio, NotebookLM, DeepSeek, Qwen, Groq Console, Perplexity, Copilot, Mistral, Hugging Face Chat, Grok (grok.com, x.com/i/grok), Cohere, Z.ai, Poe, Character.ai, Duck.ai, DuckDuckGo AI Chat, You.com, Phind, Pi, Meta AI, OpenRouter Chat.
- Download the repository — download it as a ZIP file and extract it to your local machine (or clone it with
git clone). - Open the Extensions page — navigate to
chrome://extensions/in Chrome. - Enable Developer mode — toggle the "Developer mode" switch in the top-right corner.
- Load unpacked — click the "Load unpacked" button in the top-left corner.
- Select the extension directory — browse to the extracted folder and select it.
- Done — Prompt Enhancer & RTL Fix now appears on the Extensions page. Pin it to the toolbar for quick access.
- Click the Prompt Enhancer icon in the Chrome toolbar to open the popup.
- Pick a provider from the dropdown (Groq, OpenRouter, Gemini, Cerebras, or Mistral — all have free tiers; click "Get a free key →" if you need one).
- Enter the matching API key and choose a model from the dropdown.
- Pick an Enhance style (Concise by default; also Standard / Detailed / Coding / Writing / Reasoning).
- Optionally add Custom instructions, e.g. "Always ask for a table. Prefer British English."
- Pause/resume the extension with the header switch as needed.
- Click Save Settings — the key, model, style, and instructions are stored (keys/models per provider).
-
Navigate to a supported site (ChatGPT, Claude, Gemini, DeepSeek, Perplexity, Copilot, Grok, and many more) and an Enhance button appears near the input:
-
Write your rough prompt — or just select part of the text inside the input to improve only that part.
-
Click Enhance (or press
Ctrl+Shift+Ewhile the input is focused). The button shows a spinner while processing (60s timeout):
-
The input is replaced with the enhanced prompt, and a success toast appears. Use ⧉ Copy to copy the result or ↩ Revert to step back through previous versions (up to 10 steps):
| Style | Best for |
|---|---|
| Standard | Balanced, general-purpose enhancement |
| Concise (default) | 10–15 sentences in 2–3 short paragraphs — clarifies and expands without full framework sections |
| Detailed | Acceptance criteria, edge cases, definition of done |
| Coding | Language/version, constraints, I/O examples, error handling, tests |
| Writing | Audience, tone, voice, length, structure, revision checklist |
| Reasoning | Step-by-step reasoning before the final answer |
| Action | Windows / Linux | macOS |
|---|---|---|
| Enhance current prompt | Ctrl+Shift+E |
Cmd+Shift+E |
- Provider API key required — all providers have free tiers: Groq (console.groq.com/keys), OpenRouter (openrouter.ai/keys), Gemini (aistudio.google.com), Cerebras (cloud.cerebras.ai), Mistral (console.mistral.ai). No credit card needed for any of them.
- API usage — model availability, free-tier limits, and pricing can change; check the provider's model page before use.
- Error handling — missing/invalid keys, exhausted credits, rate limits, unknown models, timeouts, and network errors are reported via an inline toast with guidance.
- Selection-aware — if text is selected inside the input, only the selection is enhanced; otherwise the whole input is used. Input is capped at ~12,000 chars per request.
- RTL fix — only leaf text nodes containing Persian/Arabic are touched (nav/buttons/inputs skipped); uses the locally bundled Vazirmatn font so it works under strict site CSPs.
Contributions are welcome! Feel free to fork this repository, make your changes, and submit a pull request.
Released under the MIT License.