From your first Telegram command to running a full server hive — Claude Academy teaches you to harness Gemi Prime, Super Claude, and Codex Max at every skill level. No jargon. Real outcomes.
What Is Claude Academy?
Claude Academy bridges the gap between "I heard AI can do that" and actually doing it. Three tracks. Three power levels. One goal — you in command of every tool in the stack.
Gemi Prime Academy starts from zero — if you can text, you can run a business AI. Learn to delegate research, calls, writing, and scheduling through Telegram in plain English.
Super Claude Academy turns your VPS into a one-person engineering firm. Manage WordPress, deploy APIs, run databases, and wire a multi-agent hive — all by Telegram.
Codex Max Academy teaches fully local AI inference. Your data never leaves your server. Swap models, run code, analyse documents, and build pipelines with zero cloud dependency.
Every module ends with something you actually ship — a running service, a published book, a deployed API, a live automation. The curriculum is built around outcomes, not theory.
Each track is layered. Beginners get clean wins fast. Intermediates learn orchestration. Advanced operators run the full hive — three agents, one command post.
Every plan includes setup by our team. You don't learn in a simulator — you learn on your own live server with your own agents. Academy content is operational, not theoretical.
Academy Tracks
Three complete curriculums — Beginner, Intermediate, and Advanced. Start anywhere. Every lesson ships a real outcome you can point to.
Gemi Prime is a fully autonomous Gemini-powered agent that lives in your Telegram. It researches, writes, calls, books appointments, publishes books, and handles your inbox — all from a chat message. This track teaches you to wield it from day one to full automation mastery.
Send your first message and understand how Gemi reasons, responds, and retains context across a full conversation thread.
→ Live conversation with your agentAsk Gemi to research any topic — competitor pricing, news, local businesses — and get a structured summary back in seconds.
→ Your first research briefGenerate proposals, emails, social captions, and job ads by describing what you need. Edit by conversation, export to PDF.
→ Sent proposal, live"Schedule a call with John for Thursday at 2pm" — Gemi handles Google Calendar, reads your inbox, and confirms everything back.
→ Appointment booked automaticallyConfigure your Gemi Voice plan. Learn how inbound calls route, how Gemi answers, and what it says to callers who reach your number.
→ First live call received by GemiFrom a one-line concept to a formatted, published book on Amazon KDP. Outline, write, proof, export, upload — Gemi runs the entire pipeline.
→ Book live on Amazon KDPScript, dial, and follow up. Gemi calls your prospect list, qualifies leads, notes responses, and sends you a daily summary with next steps.
→ First qualified lead from agent callFeed Gemi your URL. Get a full SEO audit, competitor keyword gap, and a rewritten meta deck — ready to paste into your CMS.
→ Published SEO-optimised pageContracts, NDAs, proposals, invoices, reports — describe the document and context, Gemi drafts to legal standard, exports to PDF.
→ Professional document, signed"Research the market, write the report, send it to my three clients, and calendar a follow-up for next week." One message. Done.
→ Full workflow executed autonomouslySuper Claude runs Claude Code on your own dedicated server. It edits files, deploys services, manages databases, builds APIs, and monitors the entire stack — all via Telegram. This track takes you from zero VPS experience to running a full multi-agent hive.
Understand the file system, user permissions, systemd services, and how your agents live as background processes. No terminal required yet.
→ Server mental model built"List everything running on the server" — Super Claude reads systemd status, formats it cleanly, and sends it back to you in Telegram.
→ Agent proves server access"Update the pricing on our site to $49" — Super Claude edits the page, flushes the cache, and sends you screenshot confirmation.
→ Live site change, zero dashboardDescribe what you want automated. Super Claude writes it, makes it executable, and installs the cron job — all from one Telegram message.
→ First automation running on serverConnect to your server graphically. Super Claude can operate what you see — together or fully autonomously in headless mode.
→ Visual server access confirmedDescribe an API. Super Claude writes it, tests it, deploys it behind nginx with SSL. Your endpoint is live with a real HTTPS URL.
→ Public HTTPS endpoint liveSSL certs, nginx configs, port routing, service restart on failure — Super Claude handles the entire DevOps layer via conversation.
→ New subdomain live with certCreate tables, run queries, export backups, restore from snapshot — MySQL, SQLite, and PostgreSQL, all by describing what you need.
→ DB created, queried, and backed upDrop any file in Telegram. It routes to your server hamper, gets processed by the agent, and the result comes back — automatically.
→ File processed, output deliveredWire Gemi Prime and Super Claude together. Gemi handles research and conversation. Super Claude handles execution. Learn the handoff patterns.
→ Two-agent pipeline running liveReact frontend + Python API + PostgreSQL + domain + SSL + monitoring — the entire production stack deployed in a single Claude session.
→ Production web app liveFour agents, one operator. Gemi, Super Claude, Codex Max, and a custom nanobot — wired to share context, hand off tasks, and report to one thread.
→ Full hive live and coordinatedBuild new capabilities: image search skills, CRM lookup skills, payment handlers. Package, deploy, and invoke by Telegram command.
→ Custom skill installed and workingCodex Max runs open-source AI models — Gemma, Mistral, Llama, DeepSeek — entirely on your dedicated VPS via Ollama. No API keys. No data sent to the cloud. Total inference sovereignty. This track teaches you to harness local AI for code, data, automation, and privacy-critical work.
Send a prompt, watch it route to Ollama on your VPS, receive a streamed response — zero cloud, zero API keys, zero data exposure.
→ First private AI responseGemma 3 for conversation and code. Mistral for reasoning. DeepSeek for structured output. Llama for raw capability. Pull new models in seconds.
→ 3+ models installed and testedDrop a PDF, CSV, or Word doc into Telegram. Codex reads it, summarises it, answers questions — entirely on your server, never uploaded anywhere.
→ Private document analysedDescribe a function or script. Codex writes it, runs it in a sandboxed shell, reads the output, fixes errors, delivers a working result.
→ Script written, tested, and runningBackup scripts, file organisers, log parsers — Codex handles the bash, runs it, and reports results. You describe intent; it handles terminal execution.
→ Automated shell task deployedSpin up, manage, and tear down containers by Telegram. Codex builds the Compose file, starts services, and monitors health — no Docker knowledge required.
→ Containerised app runningSchema design, migrations, query optimisation, backup schedules, integrity checks — Codex acts as your DBA, described conversationally, executed precisely.
→ Managed DB with automated backupsFeed Codex an API spec. It reverse-engineers the schema, writes a client wrapper, mocks it for testing, and documents the full integration.
→ Working API wrapper shippedCSV → cleaned → transformed → loaded into PostgreSQL → report generated. Describe the pipeline, Codex builds and schedules it end-to-end.
→ Scheduled data pipeline liveLearn which model wins at which task and switch models mid-conversation. Gemma for chat, DeepSeek for JSON, Mistral for reasoning chains.
→ Multi-model workflow builtLocal embeddings, local vector search, local LLM. Your internal knowledge base — accessible via Telegram, air-gapped from any cloud.
→ Private RAG system answering queriesRoute sensitive tasks to Codex locally, customer-facing tasks to Gemi in the cloud. Build the agent router that makes the decision automatically.
→ Intelligent routing system liveCodex writes code, writes tests, runs them, reads failures, fixes the code, repeats — autonomously. Full CI loop via Telegram, no GitHub required.
→ Self-testing service deployedServer to Engineering
Super Claude and Codex Max turn server management from a specialist skill into a Telegram conversation. Here's everything you operate — without ever opening a terminal.
nginx config, Let's Encrypt cert issuance, virtual hosts, reverse proxies, gzip, rate limiting — managed by agent command.
MySQL, PostgreSQL, SQLite. Create schemas, run migrations, backup, restore, optimise queries, and export reports — by description.
Compose files, image builds, container health, port mapping, volume mounts, rolling restarts — the agent handles the YAML.
Create, enable, disable, restart, and monitor background services. Every agent on your server runs as a managed systemd unit.
UFW rules, SSH hardening, fail2ban, audit logs, certificate rotation — described conversationally, applied precisely and verifiably.
Git hooks, build scripts, test runners, deployment targets — Super Claude wires the release pipeline from codebase to live production.
Uptime watchdogs, Telegram alerts on service failure, disk/RAM/CPU thresholds, and automated restart logic — all configured by the agent.
Automated nightly backups, on-demand snapshots, remote offload to S3-compatible storage, and restore drills — no manual scripts needed.
Real Outcomes
These aren't demos. They're production systems built by real operators using the agents you're learning. Every project here is within reach on any paid plan.
Inbound calls, live chat, appointment booking, and issue resolution — all handled by Gemi 24/7. Your customers get instant professional responses. You get a daily digest and escalations only.
voice · chat · calendar · crm · research
React frontend, Python FastAPI backend, PostgreSQL database, SSL-secured domain, Stripe payments, and admin dashboard — built, tested, and deployed from Telegram. No DevOps team needed.
react · fastapi · postgres · stripe · nginx
Connect your internal CSVs, PDFs, and databases. Codex builds a private RAG system — query your own data by Telegram, get structured answers, auto-generated reports, and trend alerts. Zero cloud exposure.
rag · vector-db · local-llm · dashboards
Gemi handles the entire KDP pipeline: title research, outline, chapter writing, back cover copy, Amazon keyword optimisation, ISBN registration, and final upload. Idea to published in hours.
kdp · writing · publishing · research · seo
WooCommerce store, product pages, full SEO audit, competitor keyword gaps, schema markup, and automated weekly content publishing — Super Claude manages the entire organic growth engine.
wordpress · woocommerce · seo · content · automation
Feed Codex your codebase. It writes unit tests, integration tests, and end-to-end specs. Runs them, flags failures by Telegram, and opens a conversation to fix them autonomously.
pytest · playwright · ci · telegram-alerts
Gemi dials your prospect list, qualifies leads with scripted conversations, handles objections, books demos, and logs every outcome to a CRM-ready spreadsheet. Human-sounding. Always on.
outbound · voice · crm · scheduling · notes
One orchestrator, four agents. Super Claude monitors all services, routes incoming tasks, deploys hotfixes, and sends you a morning briefing. Gemi handles comms. Codex handles data.
watchdog · systemd · hive · telegram · logs
Academy Enrollment
Every academy plan is $50/month — server, setup, and training included. You bring your own AI subscription. We handle everything else. Classes run 3 days a week with VIP tech support throughout.
Server + training + support · AI plan not included
Google gives new API accounts $300 in free credits — no charges until your credits run out. Based on typical Gemi Prime usage, credits last approximately 20 months. After that, estimated cost is ~$15/month on pay-as-you-go pricing. You pay Google directly — no markup from us.
Server + training + support · AI plan not included
Subscribe directly at claude.ai/upgrade. Claude Pro gives your Super Claude agent full access to Claude Sonnet and Opus models, extended context, and priority access — everything needed for server-level engineering tasks. You pay Anthropic directly.
Server + training + support · Cloud AI plan optional
Local Ollama inference is included free — no subscription needed. Optionally add OpenAI Plus at openai.com for GPT-4o access alongside your local models. Hybrid stack: private tasks go local, GPT tasks go cloud. You pay OpenAI directly.
GET IN TOUCH
Questions about Gemi Prime, VPS plans, or custom integrations? Drop us a line and we'll get back to you within a few hours.
Ready to Start?
Choose a plan, we set up your VPS and all agents, hand you the Telegram handles and RDP shortcut, and you start building from Lesson 1 on day one. No waiting. No configuration headache.