aether-mcp
v0.1.0
Published
Uncensored AI tools (chat, image, RE/malware analysis, JS deobfuscation, multi-step research) for Claude Desktop, Cursor, Cline, Zed, and any MCP client.
Maintainers
Readme
aether-mcp
Uncensored AI tools for any Model Context Protocol client — Claude Desktop, Cursor, Cline, Zed, Continue, and others.
When your main AI refuses a question, Aether answers it. Built for reverse engineers, malware analysts, security researchers, fiction writers, and anyone tired of hedged-out, watered-down responses on legal-but-edgy topics.
What's in the box
11 tools, all backed by Aether's uncensored model:
| Tool | What it does |
|---|---|
| aether_chat | Direct uncensored chat. |
| aether_explain_code | Explain code without filters — designed for IDA Pro Hex-Rays output, but works on any source. |
| aether_explain_disasm | Walk through x86/x64/ARM/MIPS disassembly. Identifies anti-debug, packers, calling conventions. |
| aether_identify_pattern | Identify what a code/asm pattern IS — VMProtect, Themida, EAC, BattleEye, custom crypto, etc. Returns structured JSON. |
| aether_suggest_bypass | Given an integrity check / anti-cheat / DRM mechanism, returns concrete bypass strategies with working code. |
| aether_engine_analyze | Game-engine-aware analysis: Unity (IL2CPP), Unreal (UObject), Source, Source 2, Godot, CryEngine, Frostbite, AnvilNext. |
| aether_protocol_dissect | Dissect a network packet capture into structured fields. For multiplayer games, malware C2, custom protocols. |
| aether_deobfuscate_js | De-obfuscate JavaScript. Identifies Obfuscator.io / JScrambler / custom obfuscators, recovers signing algorithms. |
| aether_explain_wasm | Analyze WebAssembly modules. Targets browser anti-bot (Cloudflare, Datadome, PerimeterX, Akamai). |
| aether_imagine | Unrestricted image generation. |
| aether_balance | Show your credit balance and rate-limit headroom. |
Install
# No install needed — npx fetches it on demand.
npx -y aether-mcpYou don't run this directly. Instead, your AI client launches it as a child process. Add the config below.
Get an API key
- Sign up at https://trynoguard.com — free 150 credits/day, no card.
- Open https://trynoguard.com/account → API keys → + New key.
- Copy the
ak_live_…key (shown only once).
Configure your MCP client
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"aether": {
"command": "npx",
"args": ["-y", "aether-mcp"],
"env": {
"AETHER_API_KEY": "ak_live_a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6"
}
}
}
}Restart Claude Desktop. You'll see the Aether tools in the tool palette.
Cursor
~/.cursor/mcp.json (or per-project at .cursor/mcp.json):
{
"mcpServers": {
"aether": {
"command": "npx",
"args": ["-y", "aether-mcp"],
"env": { "AETHER_API_KEY": "ak_live_a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6" }
}
}
}Cline (VS Code)
In Cline settings → MCP Servers → Edit MCP Settings:
{
"mcpServers": {
"aether": {
"command": "npx",
"args": ["-y", "aether-mcp"],
"env": { "AETHER_API_KEY": "ak_live_a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6" },
"disabled": false,
"autoApprove": []
}
}
}Zed
~/.config/zed/settings.json:
{
"context_servers": {
"aether": {
"command": {
"path": "npx",
"args": ["-y", "aether-mcp"],
"env": { "AETHER_API_KEY": "ak_live_a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6" }
}
}
}
}Continue
~/.continue/config.json:
{
"experimental": {
"modelContextProtocolServers": [
{
"transport": {
"type": "stdio",
"command": "npx",
"args": ["-y", "aether-mcp"],
"env": { "AETHER_API_KEY": "ak_live_a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6" }
}
}
]
}
}Pair with ida-pro-mcp
The combo unlocks the killer reverse-engineering workflow. Install both MCPs in Claude Desktop / Cursor:
{
"mcpServers": {
"ida": {
"command": "python",
"args": ["-m", "ida_pro_mcp"]
},
"aether": {
"command": "npx",
"args": ["-y", "aether-mcp"],
"env": { "AETHER_API_KEY": "ak_live_a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6" }
}
}
}Then in your AI chat:
"Decompile
sub_18000A4C0from IDA, identify what kind of integrity check it is, and suggest a bypass."
Your AI client calls ida.get_decompiled_function() to get the C, then aether_explain_code() and aether_suggest_bypass() to get the unfiltered analysis. Three steps, zero refusals.
Hard limits
Aether refuses only:
- CSAM
- Operational planning of real-world violence against specific identifiable people
- Synthesis instructions for weapons of mass destruction (CBRN)
- Active malware deployment against systems you do not own
Everything else — security research, malware analysis, gray-hat web work, controversial research, adult fiction, dark creative writing, blunt opinions, technical questions other AIs hedge on — Aether handles directly.
Pricing
| Plan | Per-hour rate | API keys | Daily-refresh credits | |---|---|---|---| | Free (no card) | 30 calls | 1 | 150 | | Starter ($15/mo) | 200 calls | 3 | 50 | | Pro ($35/mo) | 1,000 calls | 5 | 100 | | Unlimited ($99/mo) | unlimited | 10 | 300 |
Each call charges credits based on actual tokens used — same as the web chat. Top up anytime at https://trynoguard.com/pay with crypto (no card on file).
Privacy
- API keys are SHA-256 hashed at rest. We never see your full key after creation.
- IP and user-agent are logged on every API call for abuse detection (same policy as the web app).
- Your prompts and outputs are not sold. Conversations are not used to train third-party models.
License
MIT
