Rented Intelligence. Done Right.
~3M tokens/month free
That's 20 questions everyday
with up to ~500M tokens/month
Pay only when you scale
Install in seconds. Get grounded, source-cited research, a curated architecture knowledge engine, and frontier models — pay only when you need more, with 6-hour passes from ₹60. The freemium precursor to Expert AI Terminal Pro.
$npm i -g expert-ai
Copy
Get 10 messages and 10 MCP queries everyday - no card required. Only a free ssv.asia account
Start free, no card
Everyday research
Heavy, agentic work
Power users
For architects & AEC pros
Your hardest tasks
For teams and daily professionals, we offer flat monthly plans — bigger budgets, per-seat billing and priority support, provisioned directly. Passes are pay-as-you-go; memberships are for continuous work.
Free forever, up to 10 messages a day — grab a 6-hour pass whenever you need to keep going, or to unlock frontier models and the AEC knowledge engine.
One command installs the CLI. Get 10 messages free — no sign-up card needed.
Need more? Pick a pass right inside the terminal and pay by UPI. Each pass is a 6-hour window with its own token budget.
Buy again while active and it adds time + budget — you never lose what's left. Pay only when you actually use it.
Start free in 30 seconds · 10 messages + 10 knowledge queries daily · buy a pass from ₹60 to keep going past your free messages
No setup headaches, no API keys, no jargon. If you can copy and paste one line, you can run frontier, source-cited research from your own computer.
Expert AI runs on Node.js — the same free tool that powers most modern apps. Download it once, click through the installer, and you're set for life.
Pick the LTS version. On Linux? Install via nvm. Already have Node? Skip to step 3.
A terminal is just a box where you type commands. Use whatever you're comfortable with — it works everywhere:
Paste this and hit Enter. That's the whole install.
$npm i -g expert-ai
Copy
From any folder in any terminal, just type x and press Enter to launch Expert AI. One letter — that's your shortcut to the whole engine.
$x
Copy
On first run it opens your browser to sign in with your free ssv.asia account — once. Back in the terminal, press 1 for Research and type your first question. You're now running grounded, cited AI research. That's it. 🎉
Built for people who need answers they can trust and cite — not just confident-sounding guesses.
Type x and you're driving an agentic core, a mesh of fast expert models, and a cited knowledge base — frontier-class results at a fraction of frontier prices. Run it solo, or pair it with your Claude or Codex subscription and switch over the moment you hit their limits.
Pull answers straight from the curated AEC corpus — building codes, standards and technical references — returned with source-document citations.
Ground a reply in fresh web results — news, latest releases, current facts — summarised and cited inline, right inside your terminal.
A purpose-built agentic harness developed and maintained by ssv.asia: it doesn't just chat — it reads and edits your local files, runs commands, and loops through multi-step tasks on its own, showing its thinking as it goes.
A cascade of mixture-of-experts models that activate only the parameters each query needs — so you get frontier-class answers at a fraction of the usual speed and cost, routed automatically per task.
A curated 12M-word AEC corpus and live web search wire straight into the model — every fact comes back with a real citation you can open, not a confident guess.
Best for quick everyday questions, drafts and fast iterations.
Best for general research, writing and balanced agentic work.
Best for cited, code-accurate architecture & engineering answers.
Best for your hardest reasoning, deep analysis and long agentic runs.
Capability Benchmark is approximate, for orientation. Speed reflects industry grade throughput; prices below are effective per-million-token rates vs. public frontier list prices (input + output averaged, ₹ approx).
Blended list price per 1M tokens (input + output averaged). One flat rate — we don't charge more for output.
Bring Expert AI in the moment you hit their usage limits, or to run extra research in parallel — without burning premium quota.
Drafting, search, iteration and code exploration run on the fast free and pro routes for a fraction of frontier cost.
Deep, cited domain research (AEC) or your hardest reasoning (Foundation) — Opus / Sonnet-class quality at up to 98% less than the frontier APIs.
AI chat, AI Agents and knowledge systems built for rigorous domain expertise and sovereign data hosting.