AI can help. It can also be wrong.

Scope

These notes apply to the public Lazing website, described concepts, newsletter, community programs, creator programs, Manifestation Packs, documentation, example code and references to the local or self-hosted Lazing runtime.

No professional advice or production clearance

Lazing does not provide legal, medical, financial, tax, security, compliance or professional case-specific advice. Content, answers, manifestations, plans, decision cards, routines and workflows are starting points. They must be reviewed before real-world use.

AI errors and model limits

AI models can hallucinate, misunderstand sources, use outdated information, miss risks, infer unsafe commands, capture requirements incompletely or produce outputs that sound plausible while still being wrong. Lazing can structure, scope and trace such systems. It does not guarantee correctness, completeness, safety or suitability.

Local runtime and self-hosting

Anyone running Lazing locally, on a VPS, inside a private network or inside an organization is responsible for installation, configuration, updates, backups, access control, keys, credentials, adapters, databases, network exposure and executed commands. Before production use, permissions should be tightly scoped, data should be backed up and outputs should be reviewed.

Agents, adapters and execution

Connected tools such as Codex, Claude Code, Ollama, local models, OpenAI-compatible adapters, shell tools, browser automation or other engines may change files, process data, contact external services or prepare actions. Users must understand what an agent is allowed to do, which data is affected and what consequences could follow before execution.

Manifestation Packs and creator content

Packs, lenses, SOPs, templates, expert tracks and community contributions may come from third parties or contain domain assumptions. They are not a guarantee of any specific outcome. Creators and pack authors should describe their methods responsibly; users must check them against their context, risks and legal requirements.

No high-risk use without independent approval

Lazing is not intended to be used without separate human review in areas where errors could cause significant harm. This includes health, law, finance, human resources, critical infrastructure, safety, production systems, minors, emergency decisions and other sensitive or high-risk contexts.

Liability limitation

Use is, to the extent legally permitted, at your own risk. Liability for indirect damage, consequential damage, lost profits, data loss, business interruption, incorrect AI outputs, unreviewed automations, misconfigured adapters or commands executed by users is excluded to the extent legally permitted.

Nothing in this notice limits liability that cannot be excluded or limited under mandatory law, including liability for intent, gross negligence, injury to life, body or health, or mandatory statutory claims.

Open source and no warranty

Runtime code is provided through the project repository: https://github.com/MaximilianGerhardt/lazing. Where an open-source license applies, that license controls. Example code, documentation and previews are provided without any promise that they are error-free, secure, permanently available or fit for a particular purpose.

Responsible use

Good systems ask for boundaries. Use Lazing with small scopes, clear approvals, traceable decisions, backups, human review and special care around sensitive data. If a decision matters, a responsible human belongs in the loop.

Contact

Provider information and contact details are available in the Imprint.