Help Center

Go to /signup, enter your email, and verify it. You get 200 free credits immediately — no credit card required. You can start chatting right away.

Credits are the universal currency across all Aether surfaces. Every message, image generation, and API call costs credits based on the model and tokens used. Free accounts get 200 welcome credits + 25/month refill. You can top up anytime with crypto or card.

Core is the default model — fast, uncensored, great for most tasks. Ultra is the premium model with deeper reasoning, better at complex analysis, coding, and nuanced writing. Ultra costs more credits per message. Free users can only use Core; paid plans unlock Ultra.

No. The free tier requires only an email. When you want more credits, you can pay with crypto (BTC, ETH, USDT, 300+ coins) — no card ever needed. Card payments via Square are also available.

Aether answers the questions other AI tools refuse. Security research, reverse engineering, adult creative writing, drug harm reduction, controversial analysis, firearms knowledge — direct answers with no disclaimers, no moralizing, no refusals. Same quality, no filter.

Click the model selector in the chat composer (bottom of the screen). Core is selected by default. Ultra shows a PREMIUM badge and requires a paid plan. Your choice persists across conversations until you change it.

Click the lock icon in the chat composer. In incognito, messages aren’t saved server-side — no conversation or message rows in the database. Credits are still charged (the model ran), but nothing persists past a page reload. Good for sensitive research or one-off questions.

Yes. Both Core and Ultra support vision — drag and drop an image or click the attachment button. The model can analyze screenshots, code snippets, diagrams, photos, and any other visual content.

Never. Your conversations are never used to train, fine-tune, or improve AI models. This is a structural guarantee, not an opt-out toggle. See our trust page for the full privacy breakdown.

Core supports up to 1M tokens of context. Ultra also supports large context windows. In practice, most conversations stay well under the limit. The model sees your full conversation history within that window.

Web search is enabled by default on paid plans. The model automatically searches when it needs current information. You can also explicitly ask it to search. Free tier has limited web search access.

Memory lets the AI remember facts about you across conversations — your preferences, projects, technical stack, etc. Manage memories at /memory. Each memory is per-user and optionally per-assistant. You can view, edit, or delete any memory.

Just describe what you want in the chat. The model detects image requests and routes them to the image pipeline automatically. You can also explicitly say ‘generate an image of...’ for clarity.

Yes. Aether’s image pipeline has no NSFW filter. If the primary model’s hidden filter blocks a generation, it automatically falls back to an uncensored model. Some provider-enforced limits exist for real-person likeness editing.

Yes. Upload an image and describe the edit you want. The model routes to the image editing pipeline. For best results, be specific about what to change. Note: the model inherits the last image in a thread, so follow-up edits work without re-uploading.

Image generation costs vary by model. Fast generation (turbo) is cheaper than high-quality generation. Check your credit balance at /account. Typical generations cost 5-20 credits depending on quality settings.

A Model Context Protocol server that adds 11 Aether tools to any MCP-compatible AI client — Claude Desktop, Cursor, Cline, Zed, Continue. It lets the host AI delegate uncensored tasks to Aether when it would normally refuse. Install with: npx -y aether-mcp

An agentic coding CLI — like Claude Code, but uncensored. Run ‘aether’ for an interactive REPL or ‘aether "fix the tests"’ for one-shot mode. It reads files, writes code, runs shell commands, and connects to MCP servers. Install with: npm install -g aether-code

A browser extension (Chrome, Firefox, Edge, Brave) that adds an Aether panel to DevTools. Right-click any selection on a page to analyze it. Five modes: chat, deobfuscate JS, explain code, explain disasm, identify pattern. ~50KB, no telemetry.

Log in and go to /account. Click ‘Generate API key’. Keys start with ak_live_ and work across all surfaces — web chat, MCP, CLI, DevTools, and direct API calls. Keys are hashed after creation; you won’t see the full key again, so save it.

Yes. Install both aether-mcp and ida-pro-mcp (mrexodia/ida-pro-mcp) in your MCP client. The host AI gets Aether’s uncensored analysis tools plus IDA’s binary inspection tools in the same session — a full reverse engineering workflow.

Yes. The /api/v1/chat endpoint supports SSE streaming. Set stream: true in your request body. The API is OpenAI-compatible, so most libraries that work with OpenAI’s API work with Aether’s.

Free (200 welcome + 25/mo refill), Starter ($15/mo, 2,500 credits), Pro ($35/mo, 8,000 credits), Unlimited ($99/mo, 35,000 credits soft-cap). You can also do one-time top-ups ($5 = 1,500 credits) that never expire and stack on any plan.

Go to /pay, select a credit pack, and choose crypto checkout. We support BTC, ETH, USDT/USDC (TRC-20), and 300+ other coins via NOWPayments. No KYC required. Payments are processed in minutes.

Top-up credits never expire. Monthly plan credits refresh each billing cycle. If you cancel a plan, your remaining top-up credits stay in your account and continue working on the free tier.

When you make a purchase of $15+, your plan auto-upgrades for 30 days. $15+ → Starter, $35+ → Pro, $99+ → Unlimited. The upgrade never downgrades an existing higher plan. This means a single top-up can unlock Ultra model access.

Your first credit purchase on any provider (card or crypto) gets a 25% bonus. A $5 pack gives 1,875 credits instead of 1,500. This applies once per payment provider — so your first card purchase AND your first crypto purchase each get the bonus.

See our refund policy for details. Since credits are consumed in real-time by model inference, refunds are evaluated case-by-case for unused credits.

HSTS (1-year max-age, includeSubDomains), X-Frame-Options: DENY, X-Content-Type-Options: nosniff, Content-Security-Policy, Permissions-Policy, and Referrer-Policy. Full coverage against clickjacking, MIME sniffing, and unauthorized embedding.

Yes. Up to $2,000 per valid vulnerability report, paid in crypto. Covers the web app, API, MCP server, CLI, and DevTools extension. Responsible disclosure only — report to [email protected].

Five hard limits: CSAM, non-consensual sexual content of real people, actionable mass-casualty weapon synthesis, targeted real-person violence, and coordinated harassment. Everything else — security research, adult fiction, drug harm reduction, firearms, controversial politics, medical info — is answered directly.

Request deletion and all data — conversations, API keys, memories, payment records — is permanently removed. No 30-day waiting period, no soft-delete. Contact us via Discord or email.

We don’t train on user data, offer immediate deletion, encrypt at rest and in transit, and don’t share data with third parties. We don’t use behavioral profiling or ad targeting. See our privacy policy for the full legal framework.

Your credit balance has hit zero. Top up at /pay with crypto or card. If you’re on the free tier, you get 25 credits/month — consider upgrading or doing a one-time $5 top-up (1,500 credits, never expires).

Check that you’re sending the key as a Bearer token in the Authorization header: ‘Authorization: Bearer ak_live_...’. Keys are case-sensitive. If it still fails, generate a new key at /account — the old one may have been revoked.

1) Make sure you have Node.js 18+ installed. 2) Check your MCP config file — the command should be ‘npx’ with args [‘-y’, ‘aether-mcp’]. 3) Set AETHER_API_KEY in the env section. 4) Restart your IDE after changing config. See the setup guide for your specific client.

Run ‘aether’ without arguments — it auto-opens a browser window to grab your key on first run. If that didn’t work, set it manually: create ~/.aetherrc with your key, or set AETHER_API_KEY as an environment variable.

Aether has only 5 hard limits (CSAM, real-person NCII, mass-casualty weapons, targeted violence, coordinated harassment). If you’re hitting a refusal outside those categories, it’s likely a spurious refusal from the underlying model. Try rephrasing, or try the other model (Core vs Ultra). If it persists, report it on Discord so we can fix the system prompt.

The image pipeline has an automatic fallback chain — if the primary model’s hidden NSFW filter blocks a generation, it tries an uncensored fallback. If both fail, you’ll see an error message in chat. Try simplifying the prompt or generating in a new conversation.

Try refreshing the page. If streaming still doesn’t work, check /status for service health. Aether uses SSE streaming — some corporate proxies or VPNs can interfere. Try disabling your VPN or switching networks.

All endpoints have per-key and per-IP rate limits. Free tier has stricter limits. Wait a minute and try again, or upgrade your plan for higher limits. If you’re hitting API limits programmatically, add exponential backoff to your requests.