A Portraitor report is organised around four lenses, and the home page names all four: relationship context, Big Five (OCEAN), Transactional Analysis and change over time. Each lens lists five signals it looks at and states how its finding is written up, and almost every finding carries a confidence figure rather than a score. The public sample is labelled four of fifteen sections.<\/p>\n\n\n\n
Most tools describe what you upload. Portraitor publishes a fair amount about what comes back, lens by lens, before anyone pays. Everything below is taken from the pages Portraitor itself publishes, read today: the home page<\/a> and the privacy policy. Where those pages stop, this page stops too.<\/p>\n\n\n\n The home page sets out four lenses on the same conversation, numbered. Relationship context is the relationship it is read through. Big Five (OCEAN) covers stable traits behind the words. Transactional Analysis is described as the stance each message is sent from. Change over time is where the conversation turned.<\/p>\n\n\n\n The numbering is not decoration. Of the relationship context lens the page says it is “Established before anything else is interpreted. The same sentence means one thing from a manager and something else from a partner.” So the first lens is a precondition for the other three rather than one finding among four, and a report that got the relationship wrong would carry that error forward into everything read afterwards.<\/p>\n\n\n\n One more thing is settled before any lens runs. The home page says, of the second step, “Pick the one person you want a Portrait of. Their messages are read; everyone else stays context.” A report is about one participant, so a conversation between two people is two purchases if you want both, which is also why pointing it at your own messages<\/a> reads differently from pointing it at someone else’s.<\/p>\n\n\n\n Each lens carries a “What gets measured” panel with five signals, named in advance. That is unusual enough to be worth repeating: the categories are published before you upload anything, so the report is not free to invent a new one for your conversation.<\/p>\n\n\n\n Relationship context has five options: romantic partner, family member, close friend, colleague or manager, and mixed or unclear. The last is defined as group threads where several relationship types run at once, “Named rather than forced into one label.” The colleague entry is described as task-driven and role-bound, with authority running one way.<\/p>\n\n\n\n The Big Five lens lists openness, conscientiousness, extraversion, agreeableness and neuroticism, each written as a direction against its opposite: curiosity against a preference for the familiar, follow-through against drift and missed commitments, energy from other people against reserve, warmth against bluntness, and sensitivity to stress against an even baseline.<\/p>\n\n\n\n Transactional Analysis lists five ego states: Nurturing Parent, Critical Parent, Adult, Free Child and Adapted Child. The page reports them by situation rather than as a single verdict: “Which states dominate in calm exchanges, which appear under tension, which show up when they talk about people who are not present, and whether the balance shifts across the timeline.”<\/p>\n\n\n\n Change over time lists phases, turning points, participation, response and silence, and text markers. The fourth of those is given as “Reply speed by topic, and whether silence is cooling off, avoidance or pressure”, which is the one place the published pages commit to reading timing at all, and timing in a written thread is a harder thing to read than it looks<\/a>. The first signal, phases, divides the timeline into early, middle and late or around key dated events, which is worth knowing if you intend to upload only part of a chat<\/a>.<\/p>\n\n\n\n The reporting format is stated per lens. Relationship context gives “One label with a confidence score and a short justification, drawn from topics, terms of address and shared context, alongside an approximate positive \/ neutral \/ negative split of the whole conversation.” The Big Five lens gives a direction per trait, with how it helps and where it creates friction, and the page says each one “carries a frequency label and a confidence score”. Change over time gives a before-and-after description for each turning point, and whether the change was a short spike or became the new baseline.<\/p>\n\n\n\n The privacy policy says the same thing in its own words. Describing what the analysis produces, it lists inferences about a participant, “each with an estimate of confidence”, and adds that it also estimates the language of the conversation and whether messages appear to have been written with the help of an AI tool. The policy’s list is longer than the four lenses the home page shows: alongside personality traits it names attachment and conflict style, emotional regulation, motivations, communication style, patterns of concern as it words them, and darker personality tendencies.<\/p>\n\n\n\n The sample on the home page shows what that looks like in print. Beside an inferred relationship type it prints a confidence figure, and beside a named pattern it prints another, with the pattern itself labelled for frequency. The split of exchanges is given as ranges rather than single numbers: positive 55 to 65 per cent, neutral 20 to 30 per cent, negative or tense 10 to 20 per cent.<\/p>\n\n\n\n The sentence that defines the whole output is in the Big Five panel: “Each of the five is reported as a lean, never as a score out of ten.” A lean has a direction and a confidence attached. A score invites arithmetic, comparison between people, and a precision the written record cannot support.<\/p>\n\n\n\n That choice costs the product something. There is no number to put in a spreadsheet, nothing to compare against a friend’s result, and no total. What it buys is a report that does not assert more than its material carries, which matters because a description can feel accurate without being evidence<\/a>. The sample report’s own scope section makes the same point in one line, saying the reading is based solely on the supplied written material and that its interpretations are probabilistic.<\/p>\n\n\n\n The home page carries a readable extract of a report about a fictional participant, and labels it: “You’re seeing 4 of 15 sections.” So fifteen is the number of sections the page gives, and eleven of them are not published anywhere.<\/p>\n\n\n\n The extract opens with detected languages and a purpose statement, then runs through numbered sections: an overview of the material, scope and limitations, relationship type and emotional balance, strengths and protective behaviours, and the beginning of a section on concerning or potentially harmful patterns. The detected languages line is the visible end of the language estimate the policy mentions, which is a separate question from whether the report itself comes back in your language<\/a>.<\/p>\n\n\n\n Read against the four lenses, the extract shows how they land on the page rather than as a table of contents. Relationship type and emotional balance is the first lens. Strengths and the concerning patterns section are where the trait and ego state readings surface. Nothing in the extract is headed Big Five or Transactional Analysis, so the lenses are the reading apparatus, not the chapter titles.<\/p>\n\n\n\n Four things are genuinely not answered by any page Portraitor publishes, and it is better to say so than to guess.<\/p>\n\n\n\n Those gaps are worth holding onto when you read the output. A confidence figure printed to the nearest per cent looks like a measurement, and on the evidence of the published pages it is a statement of how firmly a reading is offered. The separate question of what the report cannot see at all, whatever its format, is covered on its own page<\/a>, and the terms set out the decisions it may not be used for<\/a> regardless of how confident a line looks.<\/p>\n\n\n\n If the format is what you were deciding on, the shape of it is now on the table: four lenses, twenty named signals, a label with a confidence figure instead of a score, and fifteen sections of which four are public. What that costs<\/a> is a separate page, and the product itself<\/a> is one upload away.<\/p>\n\n\n\n Fifteen, on the home page’s own label: the readable extract there is marked as four of fifteen sections. The other eleven are not described on any page Portraitor publishes.<\/p>\n<\/details>\n\n\n\n No. The home page says each of the Big Five traits is reported as a lean, never as a score out of ten, and gives a direction with a frequency label and a confidence score instead.<\/p>\n<\/details>\n\n\n\n A figure printed beside the finding. The privacy policy says the inferences come each with an estimate of confidence, and the extract on the home page prints one beside the inferred relationship type and another beside a named pattern. Neither page says how the figure is worked out.<\/p>\n<\/details>\n\n\n\n No. The home page says to pick the one person you want a portrait of, and that their messages are read while everyone else stays context. Two participants means two portraits.<\/p>\n<\/details>\n\n\n\n Relationship context, Big Five (OCEAN), Transactional Analysis and change over time, in that order on the home page, with relationship context established before anything else is interpreted.<\/p>\n<\/details>\n\n\n\n A format that refuses to give you a score out of ten is telling you something about how much it thinks the written record can carry.<\/p>\n\n","protected":false},"excerpt":{"rendered":" The four lenses, the twenty signals under them, and why every finding carries a confidence figure instead of a score out of ten.<\/p>\n","protected":false},"author":6,"featured_media":523,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"pp_title_highlight":"","pp_meta_title":"","pp_meta_description":"A chat analysis report is built from four lenses with five named signals each, and a confidence figure beside every label. What the published pages show.","pp_sidebar_cta":"{\"variant\": \"light\", \"heading\": \"One person per portrait\", \"bodyHtml\": \"The report reads one<\/strong> participant's messages; everyone else in the chat stays context.\", \"label\": \"Start an analysis\"}","pp_sources":"[{\"html\": \"Portraitor. Understand yourself through your conversations<\/em> (home page).\", \"href\": \"https:\/\/portraitor.ai\/\"}, {\"html\": \"Portraitor. Privacy Policy<\/em>, version 1.3.\", \"href\": \"https:\/\/portraitor.ai\/privacy\"}]","pp_endcard":"{\"h3\": \"Read the format before you read the result\", \"body\": \"Four lenses, twenty named signals, and a confidence figure beside every label. One upload, one person, one portrait.\", \"label\": \"Make your portrait\"}","pp_howto":"","footnotes":""},"categories":[3],"tags":[],"class_list":["post-522","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-understanding-yourself"],"_links":{"self":[{"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/posts\/522","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/comments?post=522"}],"version-history":[{"count":1,"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/posts\/522\/revisions"}],"predecessor-version":[{"id":524,"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/posts\/522\/revisions\/524"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/media\/523"}],"wp:attachment":[{"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/media?parent=522"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/categories?post=522"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/portraitor.ai\/blog\/wp-json\/wp\/v2\/tags?post=522"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}Four lenses, read in a fixed order<\/h2>\n\n\n\n
Five named signals under every lens<\/h2>\n\n\n\n
Every label arrives with a confidence figure<\/h2>\n\n\n\n
A lean, not a score out of ten<\/h2>\n\n\n\n
Four sections of fifteen are public<\/h2>\n\n\n\n
Where the report stops short of a number<\/h2>\n\n\n\n
Frequently asked questions<\/h2>\n\n\n\n
How many sections are in a chat analysis report?<\/summary>\n
Does a chat analysis report give a score out of ten?<\/summary>\n
What is the confidence score next to a finding?<\/summary>\n
Does the report cover both people in a conversation?<\/summary>\n
Which four lenses does a Portraitor report use?<\/summary>\n