1. Download and extract
Save the ZIP, extract it, and double-click OPEN-ME.html. A friendly welcome page guides your next steps.
Download Launch Kit →Your first step.
Your next useful build.
A clear path into AI development and agent workflows, with carefully chosen tools and projects you can understand.
Explore the learning pathDownload Launch KitFirst edition preview · September 2026 · Guide + offline Python example · No account required
What’s included in your Launch Kit
Nine files, organized to help you get started.
- Start hereA welcome page with an exercise you can try in your browser.
OPEN-ME.html - Learning guideWhat to learn first, how the example works, and what comes next.
START-HERE.md - Assistant project guideShared project instructions for your coding assistant.
PROJECT-GUIDE.md - Choose your assistantSetup paths for Codex, Claude, and other assistants.
ASSISTANT-SETUP.md - Download and setup guideSimple extraction steps and optional Python setup.
DOWNLOAD-AND-SETUP.md - Terminal shortcutsCopy-and-paste launch commands for Mac and Windows.
TERMINAL-COMMANDS.md - Practice projectA small Python example that checks data and creates a report.
workflow_demo.py - Practice dataFictional records for learning. No live market data.
fictional-markets.json - Project checksFive automated tests for the Python example.
test_workflow.py
What happens when you download?
You get these files in a ZIP. Extract it and open the welcome page. Nothing installs automatically.
What do you need?The first browser exercise needs no account or extra software. Python and your chosen assistant are only needed for later steps.
What isn’t included?A finished Kalshi bot, live trading, API credentials, or the complete planned course.
How would you like to start?
Both paths use the same kit. Choose what feels comfortable; you can switch anytime.
Guided setup is selected. No terminal is needed for the browser exercise.
Your guided path
- Download and extract the kit.
- Double-click OPEN-ME.html.
- Try the browser exercise, then follow the assistant setup guide.
For an assistant’s desktop app, open the extracted folder as a project where supported. For a chat app, follow its attachment options. The Python coding lesson still needs Python; the browser exercise does not.
One workflow before a hundred tools.
01
Define a small job
Specify the input, expected result, allowed tools, and stop condition. Start with a checklist built from three supplied notes.
02
Choose your workspace
Explore Hermes or use the coding assistant you already know. Add one skill that fits your task.
03
Run, inspect, improve
Use our included offline workflow to see validation and human review in action. Python 3.10+ is required; the example uses no paid services.
Read the complete starter guide →One kit. Your choice of assistant.
Use the same Markdown, Python, and JSON files in Codex, Claude, or another assistant with file support. The offline example also works without an assistant. Product-specific integrations have not been tested.
2. Open the welcome page
Try the first exercise directly in your browser. No installation or account needed; Python is optional for the next lesson.
Read step-by-step instructions →3. Choose your assistant
Open the folder in your preferred coding assistant and share the project guide, or run the exercise yourself.
Codex, Claude, and other assistants →Useful tools. Clear provenance.
Hermes Agent
Explore Nous Research’s agent workspace through its official setup and documentation. Provider costs depend on your configuration.
Open official documentation →Design Ops Kit
Five focused skills for frontend, dashboards, landing pages, code review, and clip editing. Choose the one your project needs.
Explore CT’s free kit →Workflow Foundations
A small Python example with fictional records, input validation, research notes, and five automated checks.
Download the example and guide →Learn the system behind the demo.
Market research workflow
Follow ingest → validate → report → review using synthetic data. Understand the structure before connecting a real service.
See the walkthrough →Kalshi prediction bot
The next proposed project: source-backed research, explicit simulation assumptions, and documented testing. Original bot code is not included in this edition.
Explore Kalshi’s demo environment →Kalshi perpetuals
A separate learning path for a distinct product and API. Margin access requires approval; no perpetuals bot is included.
Compare official API references →The included code is an offline educational scaffold. It does not connect to an exchange, simulate fills, recommend trades, or place orders.
Know exactly what you’re getting.
What is in the download?
An offline welcome page with a browser exercise, the starter guide, shared project guide, download checklist, assistant setup instructions, one Python workflow, one fictional dataset, and five tests. The example demonstrates deterministic validation and reporting; it is not yet an LLM-powered agent.
Is this the full Kalshi course?
No. This first edition establishes the learning path and an executable foundation. The original bot source, full build videos, demo integration, and production verification remain future work.
How are tools and updates handled?
Third-party projects link to their original creators. This is a dated preview without a promise of lifetime updates or ongoing support. Model services may have separate fees.
Where do the product details come from?
Official Kalshi API documentation, perpetuals access guidance, Hermes documentation, and CT’s resource directory. Reviewed September 14, 2026.
Choose your setup. Copy your command.
Extract the kit so its folder is directly inside Downloads. These commands open an already-installed assistant in the kit folder. Sign-in may be required. They do not install software or change permissions.
Paste into Terminal or PowerShell, then press Enter. If you saved the folder elsewhere, use the manual setup guide. The other-assistant option opens the welcome page; it does not install or launch an unspecified model.
Need the assistant first? Follow the official Codex CLI setup or Claude Code setup, then return here. For desktop apps, open the extracted folder using your app’s project controls.