What SageReasoning Cannot Do

We are not therapists

SageReasoning provides philosophical perspectives based on Stoic virtue ethics. It is not a mental health service, counselling tool, or therapeutic intervention. If you are experiencing psychological distress, please contact a qualified mental health professional or crisis service. Our tool cannot and should not replace that support.

The distress check runs on every surface the sweep can see. The sweep is a mechanism: it has been found structurally incomplete once, corrected, and hardened with a regression pin. It is the strongest verification we can honestly offer, not a guarantee of exhaustiveness.

If the check catches something, you receive an in-session redirect to crisis resources — and that is all that happens. Nothing is monitored afterwards, no one is notified, and there is no follow-up. We have built the detection; we have not yet built what should happen after it. If you are in crisis, please contact a crisis line or emergency services directly — do not rely on this tool to reach anyone on your behalf.

Where the AI does the reasoning — and where it does not

The practitioner tools on this site — scoring an action, the journal, the private mentor, the baseline — produce their evaluations with an AI language model (Anthropic’s Claude) working from our Stoic framework. AI models can produce outputs that are plausible-sounding but incorrect, miss important context, or reflect biases in their training data. Treat those readings as one perspective to consider, not as authoritative judgements about your character or decisions.

Our agent-facing API works differently, and it is worth being precise about it rather than describing the whole system as “AI-generated.” It runs in three layers: Claude reads your text and extracts a structured set of features from it; a deterministic, non-AI computation then scores those features — the proximity rank, the assessment of whether the act was fitting, the signed verdict — so the same features always yield the same result and the verdict can be independently reproduced; finally Claude renders that result into prose.

That middle layer removes a real source of variability, and it is the part of the system we are willing to have checked. It does not make the result infallible. The arithmetic is only ever as good as the features the first layer extracted, and a misreading there propagates silently into a confident-looking number. Nor does any layer check whether what you told us is true: the system reads how a decision was reasoned, not whether it was factually correct.

Stoicism is one philosophical tradition among many

Our reasoning framework is grounded in ancient Stoic philosophy. While we believe this tradition offers valuable insights, it is one ethical framework among many. SageReasoning does not claim that Stoic reasoning is universally correct, superior to other ethical traditions, or applicable to every situation. Other philosophical traditions (virtue ethics, deontology, consequentialism, care ethics, Ubuntu, Buddhist ethics, and many others) offer legitimate and valuable perspectives that our framework does not capture.

Stoicism emphasises the individual, not the system

Stoicism historically emphasised individual virtue over systemic critique. It has comparatively little to say about collective action, structural injustice, or the value of dissent — areas where other traditions (political philosophy, liberation ethics, critical theory, and others) offer far more. SageReasoning will help you examine your own judgements and responses; it is not a tool for analysing unjust systems, organising collective change, or deciding when principled dissent is called for. Treat its relative silence on these questions as a real limit of the framework, not as a sign that they do not matter.

Rule R19c: The framework’s known limitations must be documented and accessible.

We cannot assess your full situation

When you submit an action or decision for evaluation, you provide a text description. That description is necessarily incomplete. We cannot know your full emotional state, your relationships, your cultural context, your physical health, your financial situation, or the thousand other factors that shape a real human decision. Our evaluations work with what you tell us, which is always a simplification of reality.

Passion diagnosis is not clinical diagnosis

Our Stoic passion taxonomy identifies patterns in reasoning that ancient Stoics called “passions” (false judgements about what is good or bad). This is a philosophical framework for self-examination, not a psychological or clinical assessment. Identifying a “passion” in your reasoning does not mean you have a psychological disorder. If you find that our passion diagnoses are causing distress rather than insight, please stop using the tool and consider speaking with a professional.

Rule R20d: We specifically discourage applying the passion taxonomy to other people’s behaviour. It is designed for self-examination only.

A mirror, not a lens

SageReasoning is a mirror, not a lens: it is built for examining your own reasoning, not for diagnosing, judging, or scoring other people. Applying the framework to evaluate someone else’s character, passions, or reasoning — especially without their knowledge and consent — is a misapplication, however internally consistent the analysis may seem. Using Stoic or philosophical language to invalidate another person’s feelings or reasoning is never a legitimate use of this tool. The only reasoning you can rightly assess here is your own.

Rule R19d: The mirror principle — the framework is for examining your own reasoning, and the mentor and all tools actively discourage applying it to other people without their knowledge and consent.

Agent trust certification is not a guarantee

Sage Assent evaluates AI agents’ reasoning quality against Stoic principles. A trust certification from SageReasoning does not guarantee that an agent will behave safely, ethically, or correctly in all situations. It indicates how well the agent’s reasoning aligns with Stoic virtue at the time of assessment. Agent behaviour can change, and our assessments have the same AI limitations described above.

A configuration without Sage Reflect is not an ongoing practice

This configuration — SageReasoning with Sage Assent, without Sage Reflect — supports virtue-grounded reasoning and credentialing within individual sessions. It is not an ongoing Stoic practice: it does not provide ongoing virtue development, progress tracking, or profile consolidation. Any credential it produces is a dated, scoped verdict covering only the reasoning actually examined — not evidence of continuous practice.

Rule R19e: Where the products are offered selectively, each configuration is documented for what it supports and does not support.

We are a startup with one founder

SageReasoning is a pre-launch startup operated by a sole founder. We do not have the resources of a large organisation. Our uptime, support response times, and development pace reflect this reality. We are committed to building something principled, but we are building it honestly and incrementally.

Independence matters

Rule R20b: If you find yourself relying on SageReasoning scores to make every decision, that dependency is itself contrary to Stoic principles. The goal is to develop your own capacity for principled reasoning, not to outsource it permanently to an AI. If you notice this pattern in yourself, we encourage you to step back and practice reasoning without the tool for a while.