Analyse my own messages: what picking yourself means
You can analyse your own messages in Portraitor by choosing yourself as the person in step 2. Here is what that changes, what the portrait infers about you, and what it cannot cover.
You can analyse your own messages in Portraitor by picking your own name in step 2: your messages are read and everyone else’s stay as context. What comes back is a set of inferences about how you wrote in that one conversation, each carrying a confidence estimate that the terms call the model’s own estimate, not a measurement.
Some people export a chat because they want their own side of it described back to them, not the other person’s. Portraitor allows that, and its terms of service open the acceptable use section with the line “Portraitor is built for understanding yourself.” The process does not change when the person is you. What changes is who every inference is about, and that is worth knowing before you start.
Choosing yourself is the same click as choosing anyone else
The second of the four steps on the home page is called Who. Its instruction reads: “Pick the one person you want a Portrait of. Their messages are read; everyone else stays context.” Nothing on the page sets the person who uploaded the chat apart from anyone else in it.
The home page FAQ says the tool detects the names in your chat and offers them, and that you can pick from that list or type the name exactly as it appears. To analyse your own messages, that name is yours, spelled the way your export spells it.
The pick is fixed for that run. The same answer says that fixing a wrong choice means running it again, which uses another portrait. In a chat where two senders have similar names, check the list before you continue.
Choosing yourself does not unlock anything else. The time period in step 3 works the same way whoever you picked, from all messages down to a custom range, and the page on analysing only part of a chat covers how the person and the period combine.
Your own name is masked on your device like everyone else’s
Picking yourself does not exempt your details from masking. The privacy policy says that before any part of the conversation is transmitted, your browser removes “names, email addresses, telephone numbers, postal addresses and credentials” and replaces them with anonymous placeholders, and that this happens entirely on your own device. The policy describes no exception for the person being analysed.
The list that matches each placeholder back to a real detail never leaves your device, according to the same policy. So the analysis provider reads your messages as a placeholder’s messages, the same way it reads everyone else’s.
What masking does not hide is the subject matter. The policy is direct about it: “This removes who is speaking. It does not remove what was discussed.” If you wrote about your health, your beliefs or your relationships, those passages travel to the provider in the anonymised text, without your name attached. When the person picked is you, they are your own disclosures. The page on whether it is safe to upload your chat history goes through the rest of that journey.
What the portrait infers about the person you picked
The privacy policy lists what the analysis produces about a participant: “personality traits, attachment and conflict style, emotional regulation, motivations, communication style, patterns of concern such as manipulation or coercive control, and darker personality tendencies, each with an estimate of confidence.” When you pick yourself, every one of those categories is applied to your messages, including the last two.
The terms say the same thing from the other side: “The portrait may describe hostile, manipulative or abusive behaviour candidly, because that is what you have asked it to identify.” A portrait of yourself is not a softer version of the product. It is the same report, pointed at you.
The home page names four lenses: relationship context, the Big Five (OCEAN), Transactional Analysis, and change over time. Each Big Five trait is reported, in the page’s words, “as a lean, never as a score out of ten”, with a frequency label and a confidence score. The change over time lens includes participation, described as “Share of the talking, who starts topics, who withdraws when it matters.” For your own side, that is a description of how your share of the conversation moved.
The policy adds two estimates that are not about personality: the language of the conversation, and “whether messages appear to have been written with the help of an AI tool.” That second estimate covers the messages of the person you picked, so if you drafted some of yours with an assistant, it is your messages being looked at. How far that estimate can be taken is covered in whether anyone can tell a message was written by AI.
The portrait is delivered, the policy says, “to you alone, on screen, by email and as a copy produced on your device”, and Portraitor does not retain it.
The other side of the chat is still read, as context
“Everyone else stays context” is easy to misread as everyone else being ignored. The relationship context lens is, in the home page’s words, “Established before anything else is interpreted. The same sentence means one thing from a manager and something else from a partner.” It is reported as one label with a confidence score and a short justification, drawn from topics, terms of address and shared context.
That label frames a portrait of you. The other person’s messages are what tell the analysis what each of yours was replying to. The Transactional Analysis lens reports “Which states dominate in calm exchanges, which appear under tension”, so the portrait describes your messages as the conversation moved between the two, not your messages lifted out of it.
In a group chat the same rule holds: you are the one person, and every other member is context. The relationship lens has a label for that case, described as “Group threads where several relationship types run at once. Named rather than forced into one label.” What happens to the other members’ details is set out in what happens to everyone else in a group chat analysis.
The other people are masked exactly as you are, and the policy says neither Portraitor nor its provider learns who they are. Before you pay you still confirm that you have permission to analyse the conversation, even when the person you picked is yourself.
One conversation is not the whole of how you write
A portrait of your own messages is a portrait of your messages in one export, read through one relationship. The sample report on the home page states its scope in one sentence: “Based solely on the supplied written material.”
Run it on your messages to a sibling and then to a colleague, and the relationship lens starts from a different place each time. Nothing on the site says either of those is the real you. Each is a reading of a different record. The terms add that “The same input can produce different results at different times. Output is probabilistic rather than deterministic.” Two runs on the same chat are not guaranteed to match.
Some of what you said is not in the record at all. The policy says photographs, videos and voice recordings are never opened, and that only the written messages are read. If much of a conversation went by voice note, the portrait is built from what was typed. Messages outside the period chosen in step 3 are outside it too.
The confidence numbers need reading for what they are. The terms say: “Confidence figures shown in a portrait are the model’s own estimates, not measurements, and should be read as such.” A high figure next to a statement about you is the model’s estimate of itself, not a test result.
Reading a description of yourself carries its own trap, which is that general statements tend to feel personal. That effect has its own page, why personality descriptions feel accurate, and it is worth reading before the portrait rather than after. The wider list of what no chat analysis can reach is in what chat analysis cannot tell you.
A portrait of your own side does not answer how a particular message landed with the other person. It describes what you wrote, not anyone’s reaction to it. If that reaction is the real question, why text messages get misunderstood covers why an export cannot settle it.
Frequently asked questions
Frequently asked questions
Can I analyse my own messages instead of the other person’s?
Yes. Step 2 asks you to pick the one person you want a portrait of, and that can be you; your messages are then read and everyone else’s stay as context.
Is my name sent to the AI provider if I pick myself?
No, according to the privacy policy. Names are replaced with anonymous placeholders on your device before anything is transmitted, and the list matching placeholders back to real names never leaves your device.
Can one run cover both me and the other person?
No. Step 2 asks for one person per portrait. A portrait of the other participant means running the analysis again with their name, which uses another portrait.
What if I pick the wrong name by mistake?
The run analyses whoever was named before it started. The home page FAQ says fixing it means running it again, which uses another portrait.
Does picking myself work in a group chat?
Yes. You are the one person whose messages are read, and every other member stays as context, with their names masked on your device like yours.
Pick your own name in step 2, and read what comes back as a description of one conversation, not of you everywhere.
Read your own side of a chat
Pick your own name in step 2. Your messages are read, everyone else stays context, and every name is masked on your device first.
See how Portraitor works →