Documentation
Expert AI, documented.
Everything a much-too-busy person needs: how to sign in, activate, install, and never worry about tokens again. One letter launches it all: x.
Quickstart, from click to first prompt
- Subscribe, ₹800 / $10 per month through Razorpay (UPI, card, netbanking). Your SSV.ASIA account opens during checkout.
- Download, the success screen auto-detects your OS and offers the right 3 MB installer for Windows, macOS or Linux.
- Install, run the installer. It configures its own private runtime; no Node or npm setup required.
- Confirm, the installer opens ssv.asia in your browser. Confirm there (one click if you're already signed in), and you're in.
- Go, 500M free + 30M Pro tokens are live on your machine.
Signing in & activating
No API keys, no tokens, no config. When you run the installer for the first time it asks for your email and opens ssv.asia in your browser. You confirm the login there (sign in with your email's magic link if needed), and the CLI picks up your session. Your tier and monthly token allowance come from your SSV.ASIA membership, stored server-side.
$ expert-ai
Sign in with your email to start: you@example.com
↳ Opening ssv.asia in your browser…
✓ ssv.asia confirmed. Unlimited active. 500M free + 30M Pro ready.
Cross-platform installation
| OS | File | Size |
|---|---|---|
| Windows 10/11 (64-bit) | ExpertAI-Setup-windows-x64.exe | 3.7 MB |
| macOS Apple Silicon (M1–M4) | ExpertAI-Setup-macos-apple-silicon.pkg | 3.3 MB |
| macOS Intel (x86_64) | ExpertAI-Setup-macos-intel.pkg | 3.4 MB |
One sign-in works across all of them, install on your desktop and your laptop under the same account.
Free tokens vs Pro tokens
500M free: the everyday engine, drafting, explaining, rewriting, everyday questions. Fast, dependable, effectively unlimited at a human pace.
30M Pro: the priority lane, long code refactors, large document analysis, and heavy lifts that route through premium infrastructure with faster responses.
The monthly refresh cycle
The plan is a monthly subscription. At the start of each paid month your 500M free + 30M Pro token pools refresh in full. Unused tokens do not roll over, that monthly refresh is exactly what makes the plan effectively unlimited for uninterrupted daily work.
Cancel anytime from your ssv.asia account; your account and any already-downloaded installer remain yours.
Troubleshooting
- Sign-in not working? Confirm your membership is active and your email matches the one you paid with. Need help? Raise it on ssv.asia.
- OS asks to confirm the installer? Version 0.2.2 isn't yet Developer ID / Authenticode signed, so your OS may request confirmation. Click through; the binary is from the official Expert AI release.
- Ticked my month early? The refresh lands on your next billing date, check your account for the exact cycle.
CLI command reference
Everything you can type at the expert-ai prompt, in one table. This is the real command surface from the terminal's own help output.
| Command | What it does |
|---|---|
expert-ai | Launch the interactive REPL. |
expert-ai chat [prompt] | Start chat explicitly, optionally with an initial prompt. |
expert-ai setup | Run the guided setup wizard. |
expert-ai doctor | Diagnose config, auth, Pi runtime, and preview dependencies. |
expert-ai status | Show the current setup summary. |
expert-ai model list | List available models in auth storage. |
expert-ai model login <id> | Authenticate a model provider (OAuth or API key). |
expert-ai model logout <id> | Clear stored auth for a provider. |
expert-ai model set <provider/model> | Set the default model. |
expert-ai model tier [value] | View or set the request service tier. |
expert-ai alpha login / logout / status | Sign in to alphaXiv, clear auth, check status. |
expert-ai packages list | Show core and optional package presets. |
expert-ai packages install <preset> | Install optional packages on demand. |
expert-ai search status | Show web-access status and config path. |
expert-ai search set <provider> [api-key] | Set the web search provider. |
expert-ai search clear | Reset web search to auto while preserving keys. |
expert-ai update [package] | Update installed packages. |
--prompt "<text>" | Run one prompt and exit. |
--model <provider/model> | Force a specific model for this run. |
--thinking <level> | Set thinking level: off | minimal | low | medium | high | xhigh. |
--cwd <path> | Set working directory for tools. |
--session-dir <path> | Set session storage directory. |
--new-session | Start a new persisted session. |
Inside the REPL, slash workflows come from the live prompt-template and extension set — see the research workflows section next.
Research workflows
The terminal is built for research, not just chat. These are the pre-composed multi-step workflows you can run as one command.
| Command | What it produces |
|---|---|
deepresearch <topic> | A durable research brief with inline citations and provenance sidecar. |
lit <topic> | A literature review using paper search and primary-source synthesis. |
audit <paper> | Compares a paper's claims against its public codebase. |
replicate <paper> | Plans or executes a replication workflow. |
review <artifact> | Simulates a peer review with severity and revision plan. |
compare <topic> | A source-grounded comparison matrix. |
draft <topic> | A polished paper-style draft with equations and citations. |
autoresearch <idea> | Autonomous experiment loop — try, measure, keep what works. |
watch <topic> | A recurring research watch on a topic or company. |
code | Full-spectrum software engineering and automation. |
offline | Configure the terminal for 100% offline mode using a local model. |
lib / libsearch | Search your local knowledge vault (fuzzy RAG) without using LLM tokens. |
.provenance.md sidecar recording what was actually checked.Learn more in Research workflows, run from one command.
Expert AI vs Expert AI Terminal Pro — no confusion
Two products, one family. This site is about Expert AI — the free-to-start CLI that everyone types as x. The Pro terminal is a separate, higher-tier desktop product that we mention only so you can tell them apart.
| Expert AI (this site) | Expert AI Terminal Pro | |
|---|---|---|
| What it is | CLI in your existing terminal | Desktop / portable USB terminal |
| Launch command | x (or expert-ai) | Dedicated app / expert-ai-terminal |
| Price | Free to start · ₹800/mo unlimited · passes ₹60+ | $195 perpetual license + $36–38/mo tokens |
| Install | 3 MB installer or npm i -g expert-ai | Installer USB / desktop build |
| Best for | Everyone starting with AI in the terminal | Power users who want a dedicated, portable, air-gapped machine |
x and see a prompt, you are on Expert AI. The Pro terminal is a separate desktop app you buy intentionally — you cannot accidentally upgrade into it.Model routing & the ssv.asia lanes
Your requests are routed to the strongest available model for your plan through the ssv.asia gateway. The branded lanes map to token pools — you never need to know which lab is behind a request; you just pick a lane.
| Route | Pool it meters | Model class | Effective rate |
|---|---|---|---|
ssv.asia/free | Free tokens | Haiku + GPT class | ₹0 / M tokens · free forever |
ssv.asia/pro | Pro tokens | Sonnet class + retrieval + search | ₹30 / M tokens · ≈96% cheaper than Sonnet |
ssv.asia/aec | Pro tokens (AEC pass) | Sonnet class + cited AEC corpus | ₹30 / M tokens · flat in + out |
ssv.asia/foundation | Foundation tokens | Opus 4.8 + Fable 5 class | ₹72 / M tokens · ≈98% cheaper than Opus |
Behind the lanes: 200+ switchable models from Anthropic (Claude), OpenAI (GPT), Google (Gemini), DeepSeek, Mistral, Meta (Llama), Qwen, and more. See Frontier AI models explained for the human version.
Plans & passes at a glance
The full pricing ladder from the ssv.asia product sheet. Two shapes exist: passes (burst, 6 hours, one-off) and monthly (standing, 30-day refresh).
| Price (₹) | Product | What you get |
|---|---|---|
| ₹0 | Community | 10 free-model msgs/day + 10 AEC MCP searches/day + Quick Thoughts CLI. |
| ₹1 / query | AEC MCP | Building-code queries with citations, inside Claude/ChatGPT. Top-up ₹50–10,000. |
| ₹60 / ₹360 | Passes | 6-hour burst passes: 1M to 6M tokens depending on tier. |
| ₹800 / mo | Researcher (Unlimited) | 500M free + 30M Pro · Pro model lane · 25 MCP/day · 2 BOQ/mo · 1 device. |
| ₹1,600 / mo | Engineer | 500M free + 60M Pro · AEC engine · 75 MCP/day · 5 BOQ/mo. |
| ₹3,600 / mo | Professional | 500M free + 150M Pro · Archive Terminal + BKMS · 200 MCP/day · 2 devices. |
| ₹12,000 / mo | Founder | 500M free + 450M Pro + 3M foundation · desktop app access · 5 devices. |
| ₹45,450 / mo | Founders Desk | Services retainer — consult, install, staff. |
Separate product — Expert AI Terminal Pro (not the CLI): a $195 perpetual license + $36–38/mo token plan for the desktop / portable USB terminal. Air-gapped, zero-retention, lifetime software access. See the product demarcation table above.
Privacy & zero retention
- Local-first. Your vault, notes, and sessions stay on your machine.
- Zero data retention. The gateway treats your prompts as ephemeral; nothing is stored after the request.
- No API keys to paste. Auth is server-side, tied to your ssv.asia account.
- Air-gapped option. The
/offlineworkflow configures 100% offline use with a local model. - Portable. The Expert AI Terminal Pro USB (separate product) carries the whole installation; unplug it and nothing remains on the machine.
Keep learning — free
New to AI or the terminal? These guides are written for complete beginners, with no jargon left unexplained.