What Is What Brandon Thinks? A Guide to AI Chat Analysis

Sometimes a conversation feels easy to read until you try to explain it. One person keeps suggesting plans. Another replies warmly but rarely chooses a date. The group eventually meets, but no one remembers who made it happen. A long chat contains the details, yet finding a pattern in it can take time.

What Brandon Thinks is the name used by several AI chat analysis services. The services covered here, whatbrandonthinks.org and whatbrandonthinks.app, let you upload a supported WhatsApp or iMessage conversation and receive a structured interpretation of communication patterns. Their reports can point to message excerpts, highlight recurring dynamics and suggest questions or next steps.

The result is a second look at the messages you choose to share. It cannot tell you what someone secretly thinks or decide what a relationship means. This guide explains what the two sites offer, how to prepare a conversation and how to read a report without treating an interpretation as a fact.

Which What Brandon Thinks website is this guide about?

This article describes whatbrandonthinks.org and whatbrandonthinks.app. Both describe AI-assisted reports for WhatsApp and iMessage exports, with a free preview and a fictional sample you can inspect before uploading. Their features, purchase options and privacy terms should be checked on the site you use.

You can start a free preview on the What Brandon Thinks homepage when you are ready. Start by reading the .org sample report or the .app sample report. Both are labeled fictional, so you can see the format without sharing a real conversation.

What does What Brandon Thinks do?

The .org and .app services analyze a conversation file and turn it into a readable report. Depending on the report and available evidence, you may see:

• A short verdict: an interpretation of the overall conversation dynamic.
• Scores or signals: summaries of selected aspects of communication. These are reading aids, not clinical measurements.
• Specific findings: observations about subjects such as follow-through, recurring tension or who takes the lead in making plans.
• Message excerpts: short lines intended to show why a finding was made.
• Possible next steps: questions or actions to consider in your own words.

The .app about page describes the tool as a way to examine changes in effort and recurring patterns, with confidence labels. The .org product review says its important observations are paired with excerpts when they can be verified in the uploaded transcript. A confidence label is a cue to look at the supporting evidence; it is not a probability that an interpretation is true.

What kinds of conversations can it help you explore?

A useful question is narrow enough to check against messages. For example:

• Dating or partners: Do plans tend to be made, postponed or followed through? The chat alone cannot tell you whether someone is secretly committed or interested.
• Friends: Who starts plans, responds and takes practical steps? The chat alone cannot show the full amount of care each person gives offline.
• Family or groups: How are tasks requested, acknowledged and completed? The chat alone cannot establish everyone's intentions, responsibilities or private circumstances.

For more examples, explore the relationship chat analyzer, the friend and group chat analyzer, and the family chat guide.

For Instagram, the .org service also accepts one extracted conversation JSON file. The Instagram export guide explains how to request and find it.

How does the chat analysis process work?

The exact screens and requirements can change, so follow the instructions shown on the site you choose. The general workflow is:

1. Choose one conversation and a question. A focused question such as “How do we move from suggesting plans to confirming them?” is easier to assess than “What does this person really think of me?”
2. Prepare a supported export. Both sites describe WhatsApp ZIP or TXT and iMessage TXT inputs. The .org homepage specifies size limits; check the upload screen for the current requirements. For WhatsApp, follow the app's export instructions and choose an export without media when appropriate. iMessage may require preparing a text export; see the site's current iMessage guide.
3. Review what is in the file. Check the date range and participants. Remove material you do not want to share, while keeping enough context to understand the exchange. Editing or removing messages can affect the analysis.
4. Upload and identify participants. Confirm names and any available context so the report refers to the right people. The .app site says it analyzes the file you upload rather than requiring a live inbox connection.
5. Read the preview and inspect the evidence. Both sites advertise a personal preview before payment; their verification and access steps may differ. Check the relevant site's current terms before proceeding.
6. Decide whether to continue. If you want a paid report, review the current .org pricing or .app pricing and the checkout terms. Options and prices can change.

WhatsApp notes that certain privacy settings can prevent chat exports. Its advanced chat privacy guide explains the feature. If an export is unavailable, respect the setting rather than looking for a workaround.

How to read a report without jumping to conclusions

A report is most useful when it helps you ask a better question. Use this four-part check on one finding at a time:

1. Observation: What actually happened in the messages?
2. Evidence: Do the quoted lines support that observation when read in context?
3. Alternative: What other explanation or counterexample is plausible?
4. Next question: What could you ask or do to clarify the situation?

Consider this invented exchange, written for this article rather than taken from a customer report:

Avery: Are we still thinking about dinner Friday?

Jordan: I want to, but I need to check my shift.

Avery: Okay, let me know when you find out.

Jordan: My shift changed. Can I book Saturday instead?

Observation: Avery raised Friday; Jordan later proposed Saturday. Possible interpretation: the first plan remained open because Jordan did not yet know their schedule. Counterpoint: one exchange cannot tell us whether this is a repeated pattern. A useful next question: “Would it help if we chose a backup day when your schedule is uncertain?”

The distinction matters because words depend on relationship and social context. A 2023 ACL research paper studied how social relationships affect judgments about messages; it did not validate this particular product. Offline conversations, tone, shared history and missing messages can all change how you understand an exchange.

A short video on checking a finding

This short explainer from the report-reading guide shows why a finding should be checked against its source messages and confidence label.

Transcript: Start with one finding. Open the messages behind it. Treat confidence as a cue to review the evidence, not as certainty.

What if a finding feels wrong?

Go back to the surrounding messages, look for counterexamples and consider whether the export contains the relevant period. A score or confident-sounding sentence can still be misleading. The guide to reading a .org report walks through evidence, confidence and counterevidence in more detail.

Practical ways to use the insights

For a recurring planning problem: Look for the point where a suggestion becomes a decision. Did someone offer a specific time? Did anyone volunteer for the task? Try a concrete proposal: “Thursday at 7 or Saturday at 6? I can book once we choose.”

For an uneven-looking conversation: Compare message volume with practical contributions. A person who sends fewer messages may still arrange transport or follow through on a booking. A count alone does not measure care.

For repeated misunderstandings: Pick one example and describe the observable sequence. You might say, “When the plan stayed open after I asked twice, I felt uncertain. Can we agree on a time to confirm?” Invite the other person to correct your interpretation.

For a group chat: Look for the task that keeps falling between people, then make ownership explicit. “I can pick the restaurant; can someone else book it?” is more actionable than assigning permanent personality labels to group members.

The goal is a conversation you can have, not a verdict you need to win. If the situation involves abuse, immediate safety, legal questions or serious mental health concerns, seek appropriate human or professional support rather than relying on an AI report. For indirect written complaints, see the guide to passive-aggressive letters for ways to identify a request and respond clearly.

Five questions to take back to the conversation

A report can leave you with several interesting observations and no clear way to use them. Choose just one that you can describe without assigning a motive. These examples are starting points, not messages you need to send word for word. Adapt them to your own relationship, timing and safety.

1. When plans remain open

If a discussion contains several suggestions but no confirmed plan, ask for a choice and a decision point. This gives the other person room to say that neither option works. It also separates a scheduling problem from an assumption about interest.

Would Thursday or Saturday work, or should we pick another week?

2. When the same person keeps organizing

Look for practical contributions before assuming that one person does all the work. If planning still feels uneven, ask for a specific task instead of making a broad accusation.

I can choose a place. Would you be able to make the reservation?

3. When a short reply is hard to read

A brief answer can reflect time pressure, a different texting style or a reluctance to discuss the subject. Ask for clarification without claiming to know which explanation is correct.

I wasn't sure how to read your reply. Is this a good time to discuss it?

4. When a group task has no owner

In a group, a request addressed to everyone can be easy for everyone to overlook. Name the task and invite someone to take it on. Give people a way to decline if they cannot help.

Could someone check the opening hours by tomorrow? If not, I can do it.

5. When the report feels too certain

Write down what would challenge its claim. A report that describes one person as disengaged may have missed a later offer to help or a conversation that happened elsewhere. Looking for counterexamples is useful even when the initial interpretation feels persuasive.

The finding says: __
A message or event that might challenge it: __
What I still need to ask: __

None of these questions guarantees a particular answer. Their value is that they make a request or uncertainty visible. If the other person offers context you did not know, update your understanding rather than trying to make their answer fit the report. You can also choose not to raise a finding if the setting is unsafe or the conversation would not be helpful.

Privacy: what to check before uploading a chat

A chat export can contain another person's personal information as well as your own. Before uploading:

• Consider everyone in the conversation. Ask for permission where appropriate and avoid uploading sensitive information that others would reasonably expect to remain private.
• Inspect the export. Look for addresses, financial details, intimate messages and information about third parties. Remove what you do not need and understand that missing context can change the result.
• Read the policy for the exact domain. Check how the service processes files, whether it uses outside providers, how long data is retained and how deletion works. A promise on one site does not automatically apply to another.
• Protect the file and the report. Store the export securely, avoid sharing report links casually and delete local copies when you no longer need them.
• Check deletion controls. The .org homepage currently states that transcript deletion is scheduled after 48 hours if no report is created, or seven days after the latest report. Check the current .org safety page or .app safety page for the policy that applies to your account and report.

No privacy statement can remove the need to decide whether sharing a particular conversation is appropriate.

Frequently asked questions

Is What Brandon Thinks free?

The .org and .app sites advertise free personal previews without card details. More complete reports may require payment. Preview steps and purchase options differ, so check the current page and checkout for your chosen site.

Can it analyze WhatsApp and iMessage chats?

Both sites describe support for WhatsApp exports and iMessage text exports. File formats, size limits and preparation steps can change; check the site's upload instructions before preparing a file. The .org service also accepts one extracted Instagram or Messenger conversation JSON file. Check the upload instructions for the current limits.

Does it read someone's mind?

No. It interprets the content available in the export. It cannot know another person's thoughts or prove a motive, diagnosis or relationship outcome.

Can I see a report before uploading?

Yes. The .org fictional sample and .app fictional sample show the kind of output each site presents. Neither is a substitute for reviewing the terms of the service you choose.

Is it relationship therapy?

No. The .app describes its reports as reflection and entertainment, not professional advice. A report may help you prepare a question, but it cannot diagnose a person or replace a qualified professional.

What should I check before uploading a chat?

Review the participants, date range and sensitive information. Read the privacy and deletion policy for the exact site you choose, and consider whether the people in the conversation would reasonably expect you to share it.

Try a sample first

If you are curious, start with a fictional report and inspect the messages behind one finding. Then decide whether a real conversation is appropriate to upload, review the current privacy terms and use the free preview to assess the result. Visit whatbrandonthinks.org or whatbrandonthinks.app to explore the respective service. When you are ready, start the .org preview.