Show the evidence. Name the conflict. Leave uncertainty visible.

The conflict is part of the page

Journo researches a category in which it sells a product. That creates an unavoidable commercial conflict. Every benchmark and comparison must therefore identify Journo as the publisher, link the author, show primary evidence where it can, and state what remains unmeasured.

The standard is not “trust us because we are transparent.” The standard is enough visible evidence for a skeptical reader to reach a different conclusion.

Evidence levels

We separate four levels that marketing often blends together:

1. Direct observation: a dated result, screenshot, response, recording or product behaviour captured under a stated method.
2. First-party product fact: a feature, price or policy stated by the product’s owner, dated when reviewed.
3. Research inference: a careful interpretation of peer-reviewed or clearly labelled preprint evidence.
4. Unknown: something not tested, not disclosed or not supported closely enough to claim.

Source code can explain how Journo is intended to work. It does not prove the behaviour of the current App Store binary. Product documentation can describe a competitor. It does not prove how that product behaved in our test. We keep those statements separate.

Benchmark rules

• Use ordinary synthetic or specifically consented material, never a private customer entry.
• Publish a fixed input artifact when privacy and licensing permit.
• Record UTC date, product version, plan, device and network conditions.
• Preserve returned evidence without rewriting it.
• Publish the calculation method and machine-readable result.
• Use not measured or not tested instead of filling gaps with assumptions.
• Do not name a cross-product winner before every named product meets the same minimum protocol.
• Rerun a result after a material product change or label it stale.

Journo journal audio, transcripts, generated text, tags, feedback, raw user identifiers, provider responses and tokens must never be sent to analytics or Slack. Public research uses purpose-made non-sensitive artifacts.

Science and wellbeing claims

A plausible mechanism is not a health outcome. A study of expressive writing is not automatically evidence for speaking, transcription, an AI response or a particular app. A group average does not promise an individual benefit.

Science pages must distinguish:

• direct spoken-versus-written evidence;
• the broader expressive-writing literature;
• mechanistic work such as affect labelling;
• research on a specific clinical or study population;
• claims about the untested consumer-app bundle.

Journo is not therapy, diagnosis or medical treatment. We do not use research citations to imply otherwise.

Comparison rules

Comparison pages prefer first-party documentation, display the review date, explain who each product fits, and avoid invented hands-on claims. Paid access, affiliate compensation or supplied review access must be disclosed beside the affected claim. Journo currently has no affiliate relationship with the compared products.

Buying links, paying for positive coverage, disguising promotional pages as independent research, and acquiring expired domains to inherit their reputation are outside this policy.

Corrections

Send a correction through Journo Support with the page URL, exact statement and supporting source or evidence. We review concrete factual issues, update the page’s review date when a material correction is made, and describe a correction on-page when it changes the conclusion.

Minor spelling or formatting repairs may be made silently. We do not remove a supported result merely because it is unflattering to Journo or another product.

Who is accountable

Mehul Srivastava publishes Journo and is the named author responsible for these pages. The author profile deliberately does not claim clinical or scientific credentials. The work should stand on inspectable sources and artifacts, not borrowed authority.