SkillAgentSearch skills...

tweai

AI reply assistant for X (Twitter) — OpenAI, Grok & Gemini, 8 personas, tweet translator, smart thread context. BYOK, no subscription, open-source Chrome extension.

Install / Use

claude mcp add froggychips -- npx -y github:froggychips/tweai

If the server publishes to npm under a different name, use that package instead — check the repo README.

About this skill
🔌

MCP Server

Model Context Protocol server

Quality Score

75/100

Supported Platforms

Claude Code
Claude Desktop
Gemini CLI

Tags

Our assessment of tweai

tweai scores 75/100 on our quality scale, 852nd of 960 AI & Machine Learning skills we index.

Its MCP Server is 6.5 KB long, well organised into 16 sections with 2 code examples: a thorough specification that gives an agent plenty to work with.

It has 3 GitHub stars, so there is little community track record yet; judge it on its content.

Substance
29/30
Structure
18/20
Description
15/15
Adoption
3/20
Freshness
11/15

Maintenance, license and trust

  • The repository was last updated about 3 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.
  • Our last check on 2026-09-20 found the source still online.
  • It is released under the Apache-2.0 license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 90/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands.

Automated pattern scan on 2026-10-18. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

tweai compared with similar skills

All 4 of these similar skills score higher than tweai; compare them before choosing.

SkillScoreStarsUpdatedFormat
tweai (this skill)by froggychips7533mo agoMCP Server
claude-memby thedotmack100101.5ktodayCLAUDE.md
Agent-Reachby Panniantong10098.9k3d agoCLAUDE.md
Understand-Anythingby Egonex-AI10086.9k6d agoCLAUDE.md
headroomby headroomlabs-ai10075.7ktodayCLAUDE.md

Frequently asked questions

How do I install tweai?
Run claude mcp add froggychips -- npx -y github:froggychips/tweai. The install tabs above show the steps for each supported agent.
Which AI agents does tweai work with?
It is written for Claude Code, Claude Desktop and Gemini CLI, as a MCP Server file. Other agents that read the same format can often use it too.
Is tweai safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. It is Apache-2.0-licensed and scores 90/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is tweai still maintained?
The repository was last updated about 3 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.

TweAI — AI Reply Assistant for X (Twitter)

Open-source Chrome extension that drafts AI replies on X / Twitter in your own voice, with personas, multilingual support, and your own OpenAI key. No subscription. No backend. Your key, your data.

License: Apache 2.0 Manifest V3 BYOK

TweAI demo


Why I built this

  1. I was in Bangkok, my Thai was bad, and X's auto-translate would either break the page or render slang as nonsense. I wanted to read the timeline and reply like a human, not like Google Translate having a stroke. So TweAI started as a personal translator with persona-aware replies — long before X shipped the neural one.

Personas, flirt mode, deep author context, the careful work on prompt-injection defence — all of that is here because the original use case was actual human conversation across a language gap, not "scaling thought leadership." The extension was always for one person; it's just open-source now in case you have the same problem.


Features

  • ✍️ AI Reply with personas — 5 built-in tech-creator personas (Founder, Engineer, Researcher, Casual, Flirt) plus your own custom personas with arbitrary system prompts.
  • 🪪 Per-account memory — different default persona depending on which X account is active in the tab.
  • 🤖 Explain any tweet — 2–3 bullet summary: gist, tone, context.
  • 🌐 Auto-translate the timeline — lazy via IntersectionObserver, cached locally so you never re-translate the same tweet.
  • 🎯 Deep context — optionally include the author's last 5 tweets so the reply matches their thread.
  • 💬 DM composer — same prompt + Tweet flow for direct messages.
  • 💸 Token budget — daily soft-limit on input + output tokens; blocks API calls when exceeded.
  • 🩺 Diagnostics — one-click health check verifies key, provider connectivity, and that X DOM selectors still match.
  • 🔑 BYOK — your OpenAI (or Google Translate) key, stored locally. Nothing transits a server we control.
  • 🌍 Multilingual — UI labels in English/Russian; replies in any of 12 languages or matching the source tweet.

Install

From source (recommended while in beta)

git clone https://github.com/froggychips/tweai.git
  1. Open chrome://extensions
  2. Enable Developer mode (top-right)
  3. Click Load unpacked and select the TweAI directory

From the Chrome Web Store

Coming soon. The extension is currently in beta — install from source above.

Configure

  1. After loading the extension, click its icon → Options.
  2. Paste your OpenAI API key (sk-...). Get one at platform.openai.com.
  3. Optionally:
    • Set a default persona (Tech Founder, Engineer, AI Researcher, Casual Tech, Flirt).
    • Pick a default reply language (or auto to match your browser).
    • Switch translation provider to Google Translate API if you have a key (cheaper for high-volume timeline translation).
  4. Click Test API Key to verify, then Save.

Usage

| Trigger | What it does | |---|---| | Open x.com or twitter.com | Auto-translates each tweet as it scrolls into view (toggleable in options) | | Click 🤖 Explain under any tweet | Renders a 2–3 bullet AI analysis | | Click AI Reply with style + persona | Generates a draft reply matching the chosen voice | | Toggle Deep before AI Reply | Includes the author's recent tweets as context | | Click ✍️, type a prompt, press Enter | Sends a custom prompt to the AI for a tailored draft | | Click Tweet | Translates the draft into the source tweet's language, opens the native reply box, inserts the text |

Permissions, in plain words

| Permission | What it's actually for | |---|---| | storage | Save your API key, preferences, custom personas, per-account rules. | | scripting, activeTab | Inject the in-page UI (buttons, submenu, composer). | | webNavigation | Detect SPA navigation on x.com so the UI survives client-side routing. | | clipboardWrite | The Tweet button's clipboard fallback when the native composer isn't available. | | host_permissions | api.openai.com, x.com, twitter.com. |

Privacy

TweAI is bring-your-own-key. There is no backend. Your API key, your prompts, and the AI responses never pass through any server we control. See PRIVACY.md for the full breakdown.

Security

The extension wraps tweet text in delimiters and explicitly instructs the model to treat it as data, not instructions, defending against the most common prompt-injection attacks. To report a vulnerability, open a private security advisory. See SECURITY.md.

Roadmap

  • [x] Personas (Tech creator preset)
  • [x] BYOK + privacy-first
  • [x] Prompt-injection hardening
  • [x] Custom persona editor (write your own system prompt)
  • [x] Per-account persona memory (different voice on different X accounts)
  • [x] Token-budget and daily limits in Options
  • [x] Health-check (verify selectors and provider connectivity)
  • [ ] Chrome Web Store publication
  • [ ] Cloud-tier (24/7 monitoring, lead-gen pipeline) — for B2B; explicitly opt-in

Releasing

npm install
npm run lint
npm run format
npm run package          # → dist/ + tweai-v<version>.zip

tools/build.mjs builds the CWS-ready zip with the right inclusions (no tweai-mcp-server/, no docs/). The same script runs in .github/workflows/release.yml — pushing a v* tag uploads a GitHub release with the zip attached automatically.

CWS submission checklist: docs/CWS_REVIEW.md. Launch-day copy: docs/LAUNCH.md. Demo GIF: docs/RECORDING.md.

Contributing

See CONTRIBUTING.md for project layout, local dev, and PR conventions. Architecture overview lives in docs/ARCHITECTURE.md. If something on x.com stops working, start with docs/TROUBLESHOOTING.md.

The project has zero runtime dependencies. Source is plain MV3 JavaScript; only eslint and prettier ship as dev-deps.

Contact

  • GitHub Issues: https://github.com/froggychips/tweai/issues
  • Telegram: @froggychips

License

Apache 2.0 — see LICENSE.

Related Skills

View on GitHub
GitHub Stars3
CategoryAI
Updated3mo ago
Forks1

Languages

JavaScript

Trust signals

90/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 low1 info