Frequently Asked Questions

Yes. The full course — all 22 modules, the technique playbook, the AI agent code — is free and will stay free. No upsell, no paywall, no premium tier. The course content is licensed under CC-BY-SA 4.0, meaning you can share and adapt it as long as you credit the source and keep your version under the same licence. The code is Apache 2.0. See the licensing section below for details.

Because the techniques work, and keeping them locked behind expensive certification programmes means the people who most need them — people who can't afford a therapist, people in communities that don't have access to trained practitioners — never get them. The belief here is that a well-explained, ethically grounded free resource does more good in the world than a premium product. If you disagree, you're welcome to print this page and argue with it.

Yes. The course content is CC-BY-SA 4.0 — share it freely. The only requirements are that you attribute the source (a link back to hypnosis.mikee.ai is sufficient) and that any derivative work you publish uses the same licence. You cannot strip the licence and re-sell the content as a proprietary course. If you're building something on top of this, check the legal section.

None. Module 1 starts from the assumption that you've never seen a hypnosis demonstration outside of a TV show. If anything, less prior "knowledge" is better — most popular ideas about hypnosis (mind control, sleep, special powers) are wrong enough that they actively get in the way. The first lesson explicitly addresses this. Come curious and willing to experiment; leave the preconceptions behind.

No. This is not a certification programme and does not result in any credential. Completing the course will give you a solid working understanding of conversational hypnosis and the core therapeutic techniques, but it does not qualify you to present yourself as a certified hypnotherapist. If you want accreditation, look into the National Guild of Hypnotists, the American Society of Clinical Hypnosis, or equivalent bodies in your country. This course is excellent preparation for formal training — it's not a substitute for it.

The 22 modules represent roughly 30–40 hours of original training material. Read-through time is maybe 8–10 hours. The real time investment is practice: doing each exercise repeatedly until the pattern becomes a reflex. Igor's own guidance is to do the regression protocol 10–30 times before sitting with a real client. Budget two to four weeks of consistent practice if you want the techniques to feel natural, not scripted. There is no deadline. Go at whatever pace keeps you engaged.

You can share the material, run study groups, and teach the concepts — the CC-BY-SA licence explicitly permits this. What you cannot do is charge for a course that presents this material as your own proprietary system without attribution, or strip the licence from derivative content. If you're running a study group or teaching an informal workshop, just link back to the source. If you're building a commercial training product, read the licence carefully and make sure your derivative work is also CC-BY-SA.

The igor agent is a two-layer AI system that conducts conversational hypnotherapy sessions via text. A planner model tracks session state — which phase of the PCAT protocol you're in, what the client's problem and resource are, what technique is appropriate next. An executor model writes the actual hypnotic language: pacing statements, MBL questions, Sanctuary inductions, and so on. The result is a session that follows proper therapeutic structure while using genuinely hypnotic prose rather than generic chatbot responses. It also includes safety rails: a consent flow at the start, crisis detection mid-session, and a hard stop on dark-pattern prompts.

You need an Anthropic API key, which does require an account and is billed by usage. There is no flat monthly fee — you pay per token (roughly per word processed). A short session (10–15 minutes) typically costs well under $0.10 at current Sonnet pricing. The agent is built to use Claude Sonnet for the executor and a lighter model for the planner, so costs stay low. See the next question for a rough estimate.

A typical 15-minute session (around 2,000–4,000 tokens in and out combined) costs approximately $0.03–0.08 using Claude Sonnet. A longer 45-minute session with deeper trance work might reach $0.15–0.25. These are ballpark figures — actual cost depends on how verbose the session gets. You can monitor usage in the Anthropic console. There's no way to run up a large bill by accident unless you're running hundreds of sessions or looping without stopping.

Partially. The igor mbl command — which generates Mind-Bending Language questions — runs entirely offline from the bundled knowledge base. The igor kb-info command is also offline. However, igor chat requires an active API connection to Anthropic's servers, since it calls Claude to generate the session language. There is no local-LLM mode officially supported yet, though see the troubleshooting entry for how to attempt it.

Technically yes — igor chat will run a session without you having read anything. Practically, you'll get much more out of it if you understand what the agent is doing and why. Reading even the first three modules before your first session will let you recognise the techniques as they're applied, which deepens both the learning and the experience. The agent is a practice tool, not a replacement for the conceptual foundation.

The code licence (Apache 2.0) permits commercial use. However, deploying the agent for real clients introduces ethical and legal responsibilities that go well beyond the licence. You would need to: obtain proper informed consent, ensure your clients understand they are interacting with an AI, have a protocol for handling distress or crisis responses, and (depending on your jurisdiction) potentially hold a relevant professional qualification. Read the ethics page carefully before going anywhere near client deployment.

Single-LLM sessions tend to drift: the model gets absorbed in generating good prose and loses track of the therapeutic structure, or it stays rigidly on-script and produces prose that feels mechanical. Splitting the roles solves this. The planner's only job is to know where the session is and what should happen next — it outputs structured state, not text. The executor's only job is to write fluent, hypnotically calibrated language for the current phase. Each model can be optimised for its specific task. It's the same reason a surgeon and an anaesthetist are different people.

Yes, with effort. The planner and executor are defined as subclassable Python classes. If you want to run GPT-4o or a local Ollama model, you'd override the generate() method in each class and handle the API differences yourself. There's no official support for this, so you're on your own if something breaks. The troubleshooting page has a starting-point sketch.

Session transcripts are stored in a local SQLite database at .igor/igor.db. The agent does not automatically load previous sessions into context — each new igor chat starts fresh by default. You can pass --client-id to associate sessions with a named client, which enables future features for longitudinal work. Browsing past sessions directly is possible by querying the SQLite file with any SQLite viewer.

No. Stage hypnosis is entertainment: carefully screened volunteers, peer pressure, social performance, and often a selection process that identifies the most compliant participants before the show even starts. Conversational hypnosis — as taught in this course — is a therapeutic and communication framework. It works with the client's own goals and resources. There are no tricks, no embarrassing commands, and no "I will snap my fingers and you'll cluck like a chicken." The underlying mechanism (focused attention, bypassing habitual critical evaluation, accessing unconscious capabilities) is the same, but the application and ethics are entirely different.

Not in any therapeutically meaningful way. Hypnosis requires cooperation. The critical factor bypass works through invitation, not override — the yes-set, the pacing, the imagine-and-pretend frame all require the subject to be at least passively willing. Someone actively resisting will simply not enter the hypnotic context. This is actually a reassuring property of the technique, not a limitation. See also the ethics page on covert use.

First: "not going into trance" is almost never a binary failure. Most people experience some degree of absorbed attention, slowed thinking, or altered perception of time even on a first attempt. They just don't match their TV-derived image of "hypnosis." Second: if the response genuinely isn't happening, the HABS formula tells you exactly where to look — which of the four conditions (Hypnotic Context, Absorbed Attention, Bypass, Stimulate) is missing. Third: the MBL techniques work in a conversational context with zero formal trance. Trance depth has no correlation with therapeutic outcome. See Trance Depth in the glossary.

Neuro-Linguistic Programming and conversational hypnosis share ancestry — both trace back to Richard Bandler and John Grinder's modelling of Milton Erickson and Fritz Perls in the 1970s. NLP tends to emphasise submodality work, anchoring, and representational systems as its primary toolkit. This course emphasises the hypnotic context, affect-based state work, and the PCAT therapeutic protocol. In practice, the two overlap significantly: the law of associations, the yes-set, embedded commands, and presuppositions appear in both. If you've studied NLP, you'll recognise a lot here — and you'll also find things that go further in certain directions.

It's strongly influenced by it. Milton Erickson's core insights — that trance is a naturally occurring state, that indirect suggestion respects the client's autonomy better than direct commands, that the unconscious mind is a resource rather than a repository of drives — are woven throughout. The language patterns (embedded commands, presuppositions, metaphor) are classically Ericksonian. The course builds a more structured, drillable framework around those principles than Erickson himself ever articulated, which makes it more teachable without losing the essential spirit.

No — not without appropriate professional training and licencing. This course will give you tools that are genuinely effective for personal development, stress, motivation, and working through moderate emotional blocks. It is not a substitute for clinical training in treating depression, anxiety disorders, trauma, addiction, eating disorders, psychosis, or any other diagnosable condition. The ethics page covers this in detail, including when and how to refer. If you're already a licensed therapist, this course is a useful adjunct toolkit; if you're not, keep your scope of work firmly in the personal-development space.

Children are generally excellent hypnotic subjects — their imagination is vivid and their critical factor lighter. However, working with minors requires parental or guardian consent, additional safeguards, and (for therapeutic contexts) appropriate professional qualification. The AI agent is not designed for use with minors. If you're a qualified child therapist exploring hypnotic techniques, the course material is relevant — just apply your existing professional standards around safeguarding and consent.

The agent has a hard-coded consent flow at session start — it will not proceed without an explicit agreement from the person in the chair. Beyond the technical guardrail, the answer is the same as for any hypnotic work: covert use without consent is both ineffective (because the hypnotic context requires cooperation) and ethically prohibited. The ethics page treats this at length. Short version: don't do it, the agent won't help you do it, and even if it could, you shouldn't.

The agent includes pattern-matching for crisis language — expressions of self-harm, suicidal ideation, descriptions of abuse, and similar signals. When triggered, it exits the hypnotic frame, acknowledges what was said directly, and provides relevant crisis resources (988 in the US, 116 123 Samaritans in the UK, 13 11 14 Lifeline in Australia). It does not continue the trance-work. This response is intentionally conservative: it's better to interrupt a session unnecessarily than to miss a genuine crisis. See the troubleshooting note if the agent is triggering on casual mentions.

Self-work with these techniques is generally safe and is actually recommended as the starting point. The course explicitly suggests doing every exercise on yourself before trying it with anyone else. Self-hypnosis, the Sanctuary, and even the MBL questions can be done in self-directed mode. The main caution: if you're aware that you have a history of dissociation, trauma responses, or mental health conditions, check in with your therapist before doing regression work (Days of Wonder, the affect bridge exercises). Those techniques are specifically designed to surface older emotional material — have support in place if that's relevant to you.

An abreaction is an unexpected, intense emotional release — tears, anger, fear, or physical shaking — that arises when a technique reaches material the client wasn't consciously expecting. It can happen in the middle of what felt like a gentle session. The risk is real but manageable. The course teaches you to hold space for it (let it run if it's releasing, not overwhelming), use dissociation tools if it becomes too intense, and close every session with an educational integration period before emergence. If you're using the agent and something unexpected surfaces, type "stop" — the agent will bring the session to a calm close and give you time to reorient. If strong feelings persist after a session ends, give them space and reach out for support if needed.

Yes, if you have one. This isn't a "cover your back" suggestion — it's genuinely useful. Your therapist knows your history and can help you integrate what comes up, flag material that might be better handled in a professional session, and ensure the techniques you're practising don't work at cross-purposes to other therapeutic work you're doing. Most therapists who know anything about hypnosis will be supportive; some use similar techniques themselves. If your therapist dismisses it out of hand without engagement, that's worth a conversation too.

The igor agent source code is released under the Apache License 2.0. This means you can use it freely for personal or commercial purposes, modify it, and distribute it, as long as you include the original licence notice. You cannot use Anthropic's or Igor Ledochowski's trademarks in ways that imply endorsement. The full licence text is in the LICENSE file in the repository.

The course summaries, technique descriptions, and all written content on this site are released under Creative Commons Attribution-ShareAlike 4.0 International (CC-BY-SA 4.0). You can share and adapt the material freely — including for commercial purposes — as long as you: (1) attribute the source with a link to hypnosis.mikee.ai, and (2) release any derivative works under the same CC-BY-SA 4.0 licence. See creativecommons.org for the full terms.

Yes, within the licence terms. The Apache 2.0 code licence explicitly permits commercial use of the software. The CC-BY-SA content licence also permits commercial use, with the attribution and share-alike requirements. What you cannot do: present the course content as your own proprietary work, strip the licences from derivative content, or use Igor Ledochowski's name in a way that implies his endorsement of your product. If you're building a commercial service on top of this, you're encouraged to contribute improvements back to the open-source project.

The original training recordings and verbatim scripts are Igor Ledochowski's intellectual property and are not ours to redistribute. What this site contains are distilled summaries, principles, and technique descriptions — written in our own words from study of the course material. These are transformative works that fall within fair use for educational purposes, and they're released under CC-BY-SA. If you want the original verbatim material, Igor's work is available through his own channels.

Please do. No permission needed. A plain hyperlink to hypnosis.mikee.ai or any page within it is welcome from any context — blogs, forums, course directories, Reddit, wherever. If you want to embed content, the CC-BY-SA licence applies: attribute and share-alike. If you're a hypnosis resource directory or educational site and want a more formal description for a listing, feel free to use the description on the home page.