Home / Download

Download the agent

The complete igor Python package — knowledge base, planner + executor, safety layer, terminal & browser delivery. Free. No signup. Apache 2.0 license on code, CC-BY-SA 4.0 on content.

Get it

Direct download

Single zip with everything: source code, sanitized knowledge base, license files, README, and example .env.

Download igor-hypnosis-v0.1.0.zip

104 KB · SHA256 ad1851d6364a8664dd6ce4031d85fd1d26eba01f1cfd786a31365e59b38d187c

Let Claude install it for you

Don't want to wrangle Python, virtualenvs, and shell commands? Open Claude, drop the zip into the chat, and paste this prompt:

I downloaded this conversational hypnotherapy
agent. Please:

1. Unzip it and show me what's inside.
2. Walk me through installing it on my
   [Mac / Linux / Windows] with a clean
   virtual environment.
3. Help me set up my Anthropic API key
   in the .env file.
4. Run the smoke tests (igor kb-info,
   igor mbl "anxious" "calm").
5. Then start my first session with
   igor chat.

Stop after each step so I can confirm.

Works in Claude Code, Claude Desktop, or claude.ai. Claude knows Python, pip, every OS, and how to troubleshoot.

Or install it yourself

If you're comfortable in a terminal:

unzip igor-hypnosis-v0.1.0.zip
cd igor-hypnosis
python -m venv .venv
source .venv/bin/activate
pip install -e .
cp .env.example .env
# add your ANTHROPIC_API_KEY to .env
igor chat

See the install guide for platform-specific notes (macOS, Linux, Windows, WSL).

Quick install

From either the zip or the cloned repo:

cd igor-hypnosis
python -m venv .venv
source .venv/bin/activate            # macOS / Linux
# .venv\Scripts\activate.bat         # Windows
pip install -e .

cp .env.example .env
# edit .env: set ANTHROPIC_API_KEY=sk-ant-...

igor kb-info                          # smoke test: confirm KB loaded
igor mbl "anxious" "calm"             # offline: generate MBL questions
igor chat                             # full session in your terminal

Need more detail? See the install guide for platform-specific instructions.

What's in the package

Source code (src/igor/): The Python package itself. State machine, planner (Claude Sonnet), executor (Claude Opus), safety layer, delivery adapters (terminal + browser), SQLite memory, MBL formula generator.
Knowledge base (kb/): Three JSON files: techniques.json (37 techniques across 9 categories), patterns.json (68 language patterns), phases.json (the 7-phase session FSM). All sanitized for public release — original-voice examples; references back to the course for full sample wording.
Tests (tests/): Pytest suite — covers the MBL formula generator and the safety layer.
Documentation: README, ARCHITECTURE.md, ETHICS.md, and example .env.
License files: Apache 2.0 on code; CC-BY-SA 4.0 on knowledge base and content.

You'll need

Python 3.11+ — check with python3 --version
An Anthropic API key — sign up at console.anthropic.com. Pay-as-you-go. Each session costs roughly $1 in API usage (Sonnet planner + Opus executor, ~15 turns).
Optional: ElevenLabs API key for voice delivery (after unzipping, run pip install -e .[voice])
Optional: Playwright for browser delivery (pip install -e .[browser] && playwright install chromium)

Verify the package

The SHA256 checksum for this release is shown above. After download:

# macOS / Linux
shasum -a 256 igor-hypnosis-v0.1.0.zip

# Windows PowerShell
Get-FileHash igor-hypnosis-v0.1.0.zip -Algorithm SHA256

Compare against the published checksum on the GitHub release page or in the box above.

Licensing in one paragraph

The code is Apache 2.0: use it commercially, modify it, ship it inside your own product — as long as you preserve attribution and license notices. The knowledge base and course content are CC-BY-SA 4.0: same freedoms with the additional condition that derived content has to be shared under the same license. If you'd like to ship a commercial closed-source product that uses the knowledge base, the patterns themselves (Power Words, MBL formula structure, NAS frames, etc.) are public-domain techniques that predate this project and you can re-implement them freely from scratch.

Read this before deploying

The agent ships with a safety layer that refuses operator prompts attempting to deploy it against non-consenting people. This is intentional and non-negotiable. Hypnosis works because the unconscious is suggestible; that makes operator intent the moral pivot. The agent is for:

Self-work — running sessions on yourself
Drill and roleplay — practicing technique delivery
Consenting clients — sessions with explicit consent and clear scope
Practitioner co-pilot — a tool inside an existing licensed practice

It is not for influencing people who haven't agreed to be influenced. Read the full ethics policy before deploying anywhere user-facing.

Where this could grow

The code is Apache 2.0 — fork it, modify it, ship it. Particularly welcome directions:

Voice delivery adapter (ElevenLabs, Piper, Whisper integration)
Additional language patterns and techniques (with citations)
Roleplay mode — Igor plays the client, the user practices delivery
Internationalization — the course and prompts in other languages
Local-LLM adapter (Ollama, llama.cpp) for offline use

If you build something useful on top of this, drop a note via the email in the footer of the landing page. No formal contribution process — this is intentionally a small project.