# Admit Architect - Full Docs Corpus > The application agentic engine for MBA candidates. This file concatenates every documentation page in nav order so AI systems can ingest the complete product knowledge base in a single fetch. Source site: https://admitarchitect.com Per-page raw markdown: https://admitarchitect.com/docs/.md Index file: https://admitarchitect.com/llms.txt --- # Getting Started _Source: https://admitarchitect.com/docs/getting-started_ Create your account and run the application agentic engine end to end for the first time. Getting started with Admit Architect takes about fifteen minutes: create an account, upload your resume, run the narrative spine interview, and the agentic engine does the rest. If you only want to try things out without an account, start with our [Free Tools](/docs/free-tools). They use the same engine for a single evaluation at a time. ## 1. Create your account Admit Architect uses email and password authentication with email verification. You can also sign in with Google. After you verify your email, you will land on the dashboard. ## 2. Upload your resume The engine needs a single anchor document to seed its understanding of you. Upload the most recent version of your resume in PDF or DOCX format. Within a few seconds the engine extracts your work history, education, test scores if present, and any notable projects or awards. Your uploaded files are stored in an encrypted bucket and are never used to train any model. See [Data and Privacy](/docs/data-privacy) for details. ## 3. Complete the quick profile A short onboarding wizard collects information the resume cannot give us: target schools, target round, post-MBA goals, and a few demographic fields used by the chances prediction tool. Every field is optional but answering them lets the engine tailor every downstream step. ## 4. Run the narrative spine interview This is the single most important step. The [Narrative Spine Interview](/docs/narrative-spine) is an 8-question structured conversation that surfaces the arc of your candidacy. The engine evaluates each answer in real time for introspection depth and gently pushes back when an answer is surface-level. ## 5. Review your dean brief Once the interview is complete, the engine synthesizes a [Dean Brief](/docs/dean-brief): a narrative diagnostic with rubric scores, evidence links, and prioritized next actions. Read it carefully. It is the single source of truth every downstream step refers back to. ## 6. Open your first application From the dashboard, pick a target school and create an application. The [Essay Workbench](/docs/essay-workbench) will be pre-populated with that school's prompts, word limits, and supplementary questions. The engine already knows your spine, so Archer can draft a first pass on demand. ## How it fits in the engine Every step above is part of a single loop the engine runs and re-runs as your candidacy evolves. Resume, interview, dean brief, essays, deadlines, recommenders, and connections are not separate tools. They are stages of one orchestrated workflow, and Archer is the agent that keeps them in sync. ## Related - [The Orchestration Engine](/docs/orchestration-engine) - [Narrative Spine Interview](/docs/narrative-spine) - [Dean Brief](/docs/dean-brief) --- # The Orchestration Engine _Source: https://admitarchitect.com/docs/orchestration-engine_ How the application agentic engine plans, calls tools, and writes across every stage of your MBA application. The orchestration engine is the core of Admit Architect: an agentic platform that runs a continuous Ingest, Interrogate, Synthesize, Draft, Orchestrate, Submit loop across your entire MBA application. Most admissions tools are point solutions. A resume reviewer. An essay grader. A chatbot. Admit Architect is different. It is an admissions operating system where every stage of your candidacy is connected to every other, and an agent named Archer plans and executes multi-step work across all of them. ## The six stages ### 1. Ingest The engine pulls in raw material: your resume, transcripts, test scores, demographic facts, and anything else you attach. Files are parsed into structured text. Every extracted claim becomes a Fact in your [Profile Memory](/docs/profile-memory), linked back to an Evidence Span inside the source document. ### 2. Interrogate The engine runs the [Narrative Spine Interview](/docs/narrative-spine), an 8-question structured conversation that asks about origin, arc, leadership, values, goals, why MBA, why now, and why school. Each answer is evaluated in real time for introspection depth, and the engine surfaces targeted follow-ups rather than accepting surface-level responses. ### 3. Synthesize Once the interview has enough signal, the engine synthesizes a [Dean Brief](/docs/dean-brief): a narrative diagnostic with rubric scores, per-dimension feedback, evidence links, and prioritized next actions. The brief is append-only versioned, so every refinement over time is traceable. ### 4. Draft With your spine in hand, the engine drafts essays in the [Essay Workbench](/docs/essay-workbench). Each draft is grounded in the spine, pulls from evidence, and respects school-specific prompts and word limits. You can ask Archer for suggestions, alternate angles, or outright rewrites on demand. ### 5. Orchestrate The engine manages the parts of an application that are not essays: [Deadlines](/docs/deadlines-portfolio), [Recommenders](/docs/recommenders), and [Connections](/docs/connections). It flags at-risk work, chases recommender invitations, and drafts school-specific outreach to alumni and current students. ### 6. Submit When a school application is complete, the engine validates it against the school template, confirms every component is present, and hands it off to you for final review. After submission, it continues monitoring status and feeds any new information back into the loop for your next school. ## What makes it agentic Three things distinguish the engine from a conventional chatbot or point tool. **Multi-step planning.** Archer does not just answer a question. It can read your spine, load a school template, diff your current draft against that template, and propose a concrete edit, all in one turn. Every action is an explicit tool call the engine logs, so you can always audit what it did. **Propose-then-apply writes.** Any change Archer wants to make to your application is returned as a proposal. You review the diff and click Apply when you are ready. Read tools run freely, write tools always go through your approval. See [Mind Sessions](/docs/mind-sessions). **Persistent memory.** Archer remembers. Every fact, every evidence span, every interview answer, every dean brief version, every essay revision is stored and re-used across sessions. You never have to re-explain yourself, and the engine only gets sharper as you add material. ## How it fits in the engine This page is the fits-in-the-engine page. Every other doc in this guide describes one stage or one surface of the orchestration loop. When you are ever confused about where a feature sits, come back here and find the stage it belongs to: Ingest, Interrogate, Synthesize, Draft, Orchestrate, or Submit. ## Related - [Getting Started](/docs/getting-started) - [Narrative Spine Interview](/docs/narrative-spine) - [Mind Sessions with Archer](/docs/mind-sessions) - [Profile Memory](/docs/profile-memory) --- # Narrative Spine Interview _Source: https://admitarchitect.com/docs/narrative-spine_ An 8-question structured interview that surfaces the arc of your MBA candidacy and feeds every downstream stage of the engine. The Narrative Spine Interview is an 8-question structured interview that discovers the arc of your MBA candidacy and seeds every other stage of the agentic engine. Most admissions essays go sideways because the candidate has not yet answered, in their own words, the questions every adcom is actually asking. The spine interview is how Admit Architect forces those questions to the front of the process, before a single essay is written. ## The eight questions The interview walks through eight dimensions in order. Each one is required to move on, but you can revisit and refine any answer at any point. 1. **Origin.** Where does your story begin? What context shapes how you see the world? 2. **Arc.** What is the through line of your career so far? What pattern connects your choices? 3. **Leadership.** When have you actually led, and what did that feel like in the moments that mattered? 4. **Values.** What do you refuse to compromise on? What are you willing to be unpopular for? 5. **Goals.** What do you want to be doing in ten years, and why does that specific thing matter to you? 6. **Why MBA.** What do you need from an MBA that you cannot get any other way? 7. **Why now.** Why is this the right year in your life for graduate school? 8. **Why school.** For each target school, why is it the one school that actually fits your arc? ## How the engine evaluates answers Every answer is read by the engine in real time and evaluated across five dimensions of introspection depth: specificity, honesty, causal reasoning, self-awareness, and coherence with your prior answers. The engine does not give you a numeric score. Instead, it tells you what kind of follow-up your answer needs and offers a coach-like nudge if you have stayed at the surface. You can write a short draft answer and ask the engine to probe it. It will reply with the specific things your answer is still missing, in plain language, before you commit. ## Append-only and re-runnable Every answer you give is stored as a versioned turn in an Interview Session. You can reopen the interview at any point, edit an answer, or restart a dimension from scratch. Previous versions are never lost, so Archer can always refer back to the original for context. ## How it fits in the engine The spine interview is the Interrogate stage of the orchestration loop. The answers become the raw input for the Synthesize stage, which produces your [Dean Brief](/docs/dean-brief). Every downstream stage, essay drafting, connections, recommender coordination, refers back to the spine, so the quality of this one step sets a ceiling on everything else. If the spine is vague, every essay will be vague. If the spine is specific and honest, every essay can be specific and honest too. ## Related - [Dean Brief](/docs/dean-brief) - [Narrative Flow](/docs/narrative-flow) - [The Orchestration Engine](/docs/orchestration-engine) --- # Dean Brief _Source: https://admitarchitect.com/docs/dean-brief_ A narrative diagnostic that synthesizes your spine interview into rubric scores, evidence-linked feedback, and prioritized next actions. The Dean Brief is a narrative diagnostic that synthesizes your Narrative Spine Interview into a rubric-scored evaluation, evidence-linked commentary, and a ranked list of next actions. It is the single most important artifact the agentic engine produces. Every downstream stage, essay drafting, connections, recommender coordination, refers back to the current brief as the source of truth about your candidacy. ## What the brief contains A Dean Brief has four sections. ### Narrative spine summary A two to three paragraph retelling of your arc in the voice an admissions committee would use. This is not your self-description. It is the engine's reading of you, built from your interview answers and profile facts. ### Rubric evaluation A per-dimension evaluation across the core axes an adcom actually cares about: narrative clarity, leadership pattern, goal credibility, fit, and authenticity. Each dimension includes a written assessment and a relative rating you can use to see where you are strong and where you are thin. The engine does not publish the exact weights or formulas it uses to produce a rating. The goal of the brief is to point you at what to fix next, not to give you a grade you can argue with. ### Evidence links Every claim in the brief is linked back to the exact answer or resume span it came from. You can click any citation and jump to the underlying evidence in your [Profile Memory](/docs/profile-memory) or your interview answers. If the brief says something you do not recognize, you can audit it. ### Next actions A ranked list of the highest-leverage things you can do right now. These are not generic tips. They are specific prompts for specific gaps the engine noticed, like "your leadership examples are all reactive" or "the why-MBA answer does not yet connect to the goal you stated". ## How the brief is generated The brief is produced by a staged synthesis pipeline, not a single call. The dashboard shows a stepped progress indicator while it runs so you can see the engine work through your spine, score the rubric, resolve evidence links, and rank actions. If any stage fails, the engine stops and surfaces the error; it never fabricates a partial result. ## Append-only versioning Every regeneration of the brief is saved as a new version. Previous versions are never overwritten. This matters for two reasons: you can watch your narrative mature over time, and Archer can compare the current version against earlier ones to tell you what changed. You can regenerate the brief at any time, but in practice you only need to after you have added meaningful new material (a new answer, a new resume version, a new interview revision). ## How it fits in the engine The brief is the output of the Synthesize stage of the [orchestration engine](/docs/orchestration-engine). It takes the spine interview as input and becomes the anchor every later stage reads from. When you draft essays, Archer reads the latest brief. When you coordinate recommenders, Archer reads the brief. When you draft a connection outreach, Archer reads the brief. Keep it fresh and every other stage gets sharper. ## Related - [Narrative Spine Interview](/docs/narrative-spine) - [Narrative Flow](/docs/narrative-flow) - [Essay Workbench](/docs/essay-workbench) - [Profile Memory](/docs/profile-memory) --- # Narrative Flow _Source: https://admitarchitect.com/docs/narrative-flow_ A Sankey visualization that maps your identity and evidence domains into essays across your school portfolio. Narrative Flow is a Sankey visualization that shows how the pieces of your narrative spine flow into the essays of every school in your portfolio. It answers a question every serious applicant eventually asks: am I actually telling a consistent story across all of my schools, or am I repeating the same three anecdotes and hoping nobody notices? ## What it shows The diagram has three columns. **Left column: identity.** A small set of high-level identity elements the engine extracted from your spine and resume. These are things like origin, values, professional domain, community, and goals. **Middle column: domains.** The concrete evidence domains your candidacy pulls from, such as specific projects, employers, community work, or life events. Each domain has a weight based on how central it is to your overall story. **Right column: essays.** Every essay you are currently drafting, per school. A ribbon connects each domain to the essays it is currently cited in, with width proportional to how heavily that essay leans on that domain. ## What to look for Three patterns are worth reading the diagram for. **Overdependence.** If a single domain feeds most of your essays, you may be telling the same story six ways. That is not always wrong, but you should decide intentionally rather than by accident. **Orphans.** If a domain has zero outgoing ribbons, you have evidence the engine knows about that you are not yet using. That might be a deliberate choice, or it might be a missed opportunity. **Gaps.** If an essay has very few incoming ribbons, it is leaning on thin evidence and will probably read generic. Ask Archer "is my narrative flow balanced" in a [Mind session](/docs/mind-sessions) and the engine will read this diagram plus your current drafts and propose specific swaps. ## How it fits in the engine Narrative Flow sits between Synthesize and Draft in the orchestration loop. The [Dean Brief](/docs/dean-brief) tells you how your spine reads to an adcom. Narrative Flow tells you how that spine is being distributed across your actual essays. You use it to spot imbalance before it becomes a pattern, then fix it in the [Essay Workbench](/docs/essay-workbench). ## Related - [Dean Brief](/docs/dean-brief) - [Essay Workbench](/docs/essay-workbench) - [Profile Memory](/docs/profile-memory) --- # Essay Workbench _Source: https://admitarchitect.com/docs/essay-workbench_ A per-school essay editor with evidence links, append-only versioning, and agent-driven suggestions grounded in your narrative spine. The Essay Workbench is a per-school essay editor that is pre-wired with school templates, grounded in your narrative spine, and connected to Archer for evidence-aware suggestions you can apply in one click. It is where most of your writing time will actually be spent. ## Template-aware When you create an application for a school, the workbench loads that school's template: every required and optional essay, every short answer, every supplementary question, the word limit or character limit on each, and the prompt text as the school published it. You never type a prompt or a limit yourself, and you never need to remember that school X counts characters while school Y counts words. Admit Architect ships templates for 20 top MBA programs in the current cycle. See [Deadlines and Portfolio](/docs/deadlines-portfolio) for the full list. ## Append-only versioning Every save creates a new version of the submission. Previous versions are never deleted, and you can scroll through the history of any essay to see how it matured. This matters when you are comparing two directions for the same prompt or when you want to revert a change Archer suggested. ## Agent-driven suggestions You can ask Archer for help inside the workbench at any time: "tighten this", "the opening feels generic", "am I answering the prompt", "does this match my spine". Archer reads the current draft, your spine, your dean brief, and the school's template, and returns a response grounded in all of them. Writes are always propose-then-apply. Archer never silently overwrites your draft. Instead it returns a proposal, you see the diff, and you click Apply when you are satisfied. ## Evidence links Any sentence in your essay can be linked to one or more evidence spans in your [Profile Memory](/docs/profile-memory). The engine uses those links to double-check that your claims are supported, to surface matching evidence when you start a new draft, and to warn you if you are making a claim the engine has no evidence for yet. ## How it fits in the engine The workbench is the Draft stage of the orchestration loop. It reads from Synthesize (your [Dean Brief](/docs/dean-brief) and spine) and writes to Orchestrate (your [Deadlines](/docs/deadlines-portfolio) and completion state). If you ever feel lost inside the workbench, the right move is to step back and run [Narrative Flow](/docs/narrative-flow) to see how this essay sits in your overall portfolio. ## Related - [Narrative Spine Interview](/docs/narrative-spine) - [Dean Brief](/docs/dean-brief) - [Narrative Flow](/docs/narrative-flow) - [Mind Sessions with Archer](/docs/mind-sessions) --- # Recommenders _Source: https://admitarchitect.com/docs/recommenders_ Invite, track, and coordinate recommenders across every school in your portfolio from a single place. The Recommenders surface lets you invite letter writers, track their progress across every school you are applying to, and hand them school-specific question guidance the engine wrote based on your narrative spine. Coordinating two or three recommenders across six or eight schools is the part of the MBA application most candidates underestimate. Admit Architect treats it as a first-class stage of the orchestration loop. ## Invite and track For each application, you can invite a recommender by name and email. The engine sends them a secure link they can open without creating an account. You can see the status per recommender per school at a glance: Invited, In Progress, or Submitted. If a school asks for a specific answer format (box ratings, short answers, free-form letter), the workbench shows the recommender exactly what the school expects, so nobody has to guess. ## School-specific guidance When a recommender opens their link, they do not see a blank page. They see the school's actual prompts, the word limit or character limit for each, and a short "Archer Tip" for each question that gives them strategic context: what this school is really asking for, what strong answers tend to cover, and common pitfalls to avoid. The tips are not a script. They are a briefing. The recommender still writes in their own voice, and their draft is never shown to you or auto-edited by the engine. You can still forward a recommender your resume and your dean brief summary if they ask for more context. The engine does not do that for you, because the decision of what to share with a letter writer is yours. ## Reminders and chasing The engine tracks submission deadlines per school and will surface a reminder on your dashboard if a recommender is still outstanding as a deadline approaches. You can trigger a reminder email yourself, and the engine will never send anything to a recommender without your explicit action. ## How it fits in the engine Recommenders are part of the Orchestrate stage. The engine reads from your [Deadlines and Portfolio](/docs/deadlines-portfolio) to know when to chase, and from your narrative spine and [Dean Brief](/docs/dean-brief) to generate the per-school guidance your recommenders see. If the spine or the brief changes materially, the guidance updates the next time you open it. ## Related - [Deadlines and Portfolio](/docs/deadlines-portfolio) - [Dean Brief](/docs/dean-brief) - [Mind Sessions with Archer](/docs/mind-sessions) --- # Deadlines and Portfolio _Source: https://admitarchitect.com/docs/deadlines-portfolio_ A multi-school timeline of rounds and deadlines with at-risk flagging across your entire application portfolio. Deadlines and Portfolio is a unified timeline of every school, round, and component in your application, with at-risk flagging tied to how much work is actually left on each one. Instead of juggling six or eight school calendars by hand, you get one view of every deadline, every component, and what the engine thinks you should be working on right now. ## Multi-school calendar When you add an application, the engine loads that school's current-cycle deadlines for each round. You see the full list in calendar order: application deadline, recommendation deadline, decision release, interview window, and any supplementary events the school publishes. You can switch rounds per school at any time. The engine will not pressure you into a round that no longer fits your readiness. ## At-risk flagging Each application has a completion state the engine derives from your actual submissions: how many essays are done, how many are in draft, whether your recommenders are in motion, whether the dean brief is fresh. If that state falls behind your chosen round's deadline, the engine flags the application as at risk on the dashboard and in the timeline view. At-risk is not a judgment. It is a signal. You can decide the engine is wrong about an application and snooze the flag. The point is to surface it so you can decide. ## Timeline context Your overall timeline is also a first-class field. You can tell the engine whether you are in a calm month, a busy quarter, or a deadline crunch. Archer uses that context to decide how aggressive its next-action suggestions should be. In a calm week it might suggest running a second pass on a spine answer. In a crunch week it will stay focused on unblocking the nearest deadline. ## How it fits in the engine Deadlines and Portfolio are the clock for the Orchestrate stage. Archer reads it constantly: to prioritize next actions on the dashboard, to decide when to chase [Recommenders](/docs/recommenders), to warn you when [Essay Workbench](/docs/essay-workbench) work is slipping. If you ever feel like the engine is nagging you about the wrong thing, check your round choices and your timeline context here first. ## Related - [Recommenders](/docs/recommenders) - [Essay Workbench](/docs/essay-workbench) - [The Orchestration Engine](/docs/orchestration-engine) --- # Connections _Source: https://admitarchitect.com/docs/connections_ School-specific contact discovery and outreach drafting grounded in your narrative spine. Connections is the engine's contact discovery and outreach layer: it helps you find the right alumni and current students for each of your target schools and drafts school-specific, spine-aligned outreach you can send from your own account. Networking is the part of an MBA application that rewards candidates who are organized and punishes candidates who are not. Admit Architect treats it as another stage of the same orchestration loop. ## How contacts show up You bring the contacts. Two paths are supported: paste a LinkedIn URL or upload a CSV export of a list you already keep. The engine does not scrape and does not broker anonymous networking introductions. It is a tool for making the people you already have access to more useful, not a replacement for actually having a network. Once a contact is in the system, the engine enriches what you know about them: school, graduation year, function, industry, and shared background with you where public info supports it. ## Match reasoning For each target school, the engine scores your contacts along multiple dimensions: shared school, shared employer, shared function, shared geography, shared community, and how well their career arc maps to your goals. The matches are shown with the reasoning behind them, so you can sanity-check before you spend a message on a lukewarm match. A strong match is not the person with the most overlap. It is the person whose own path mirrors the specific next step you want to take. Ask Archer in a [Mind session](/docs/mind-sessions) if you are unsure who to reach out to first. ## Outreach drafting When you open a contact, Archer can draft a personalized outreach message. The draft pulls from your narrative spine, cites specific things the contact has in common with you, and respects the short-message norms of professional outreach. Every draft is a proposal you review and edit before anything is sent. The engine never sends on your behalf. ## How it fits in the engine Connections are part of the Orchestrate stage. The engine reads your [Dean Brief](/docs/dean-brief) and [Profile Memory](/docs/profile-memory) to know what to emphasize in outreach, and the [Deadlines](/docs/deadlines-portfolio) surface to know how much networking time you actually have. It stays out of your way in a crunch week and leans in during a calm one. ## Related - [Dean Brief](/docs/dean-brief) - [Mind Sessions with Archer](/docs/mind-sessions) - [Deadlines and Portfolio](/docs/deadlines-portfolio) --- # Chances Prediction _Source: https://admitarchitect.com/docs/chances-prediction_ Fit scoring and admission-chance prediction across 40+ MBA programs, updated as your candidacy matures. Chances Prediction scores your candidacy across 40+ MBA programs and classifies each school as reach, target, or safety, using the same signals an admissions committee actually weighs. The scoring runs as a stage of the engine so your predictions refresh as you improve your spine, your resume, or your test scores. They are not a one-shot calculator you run once and forget. ## What the prediction uses The engine pulls five broad signal categories from your profile: your academic record, your work experience, your test scores if present, your narrative strength as read from your spine and dean brief, and demographic context. It compares those signals against the most recent class profile the engine has for each program. The output is a numeric fit score per school, a classification (reach, target, or safety), and a short list of the specific things most limiting your current score. Admit Architect does not publish the exact weights it uses for each signal. The score is meant to help you prioritize, not to give you a number you can game. ## Why predictions move over time Every update to your spine interview, your resume, or your test scores can shift the score for a school. A candidate who walks in with a blank profile and finishes a complete spine interview can see meaningful movement purely from the engine having a clearer read on their narrative. That is the point. The prediction is not a static grade. It is a moving diagnostic that rewards the work you put into the loop. ## Public Chances Calculator A simpler, single-run version of this tool is available at [admitarchitect.com/tools/admission-chances](https://admitarchitect.com/tools/admission-chances) with no account required. It uses the same scoring engine but does not persist your profile or update as you change it. See [Free Tools](/docs/free-tools). ## How it fits in the engine Chances Prediction is part of Orchestrate. It does not change your narrative, but it shapes how you spend your time across schools. If an application is classified reach, the engine will usually suggest higher investment in your [Essay Workbench](/docs/essay-workbench) drafts for it, and more rigorous [Connections](/docs/connections) outreach to the specific alumni profile that correlates with historical admits. ## Related - [Free Tools](/docs/free-tools) - [Deadlines and Portfolio](/docs/deadlines-portfolio) - [Dean Brief](/docs/dean-brief) --- # Mind Sessions with Archer _Source: https://admitarchitect.com/docs/mind-sessions_ The Archer agent panel that reads your application state and proposes changes you can apply in one click. Mind Sessions are the surface where you work directly with Archer: a floating agent panel that can read any part of your application and return proposals you can apply with one click. If the orchestration engine is the circulatory system, Mind sessions are the interface you use when you want to drive something specific. ## Read freely, write by proposal The core safety rule of Mind sessions is simple: Archer can read anything in your application without asking, but it can never silently change anything. When Archer decides a change should be made, it returns the change as a proposal. You see the full diff, the reason behind the proposal, and an Apply button. Nothing actually changes until you click Apply. This is true for essays, for facts, for narrative elements, and for deadlines. Two actions are exceptions and are applied immediately because they are additive and safe: recording a new fact about yourself and marking a deadline complete. Everything else is proposal-first. ## Persistent, per-session memory Each Mind session is persistent. Open the panel, ask a question, close the panel, and come back the next day. The session is exactly where you left it. You can start a new session when you want a clean context, or continue a session when you want continuity. Archer also reads from your broader [Profile Memory](/docs/profile-memory) in every session, so it never needs to be re-told the basics about you. ## Multi-step work Archer does not just reply with a single answer. A single turn can involve planning, reading several parts of your application, calling multiple tools, and then returning a final response with a proposal attached. You can watch the plan and the tool calls in the panel in real time. Every tool call and every proposal Archer makes is logged as an audit trail on the session. If something changes and you do not remember why, you can scroll back and see exactly what Archer did. ## How it fits in the engine Mind sessions are the interactive surface on top of the [orchestration engine](/docs/orchestration-engine). The engine runs on its own schedule to surface next actions and keep your portfolio moving, but when you want to drive something manually ("rewrite this paragraph in my voice", "tell me why this essay feels thin", "draft an outreach to this alum"), Mind is where you do it. ## Related - [Profile Memory](/docs/profile-memory) - [The Orchestration Engine](/docs/orchestration-engine) - [Essay Workbench](/docs/essay-workbench) --- # Profile Memory _Source: https://admitarchitect.com/docs/profile-memory_ Archer's persistent memory layer of facts, evidence spans, and evidence links across every document and interview in your account. Profile Memory is the persistent layer Archer uses to remember everything about you: the facts it has learned, the exact evidence in your documents that supports each fact, and the links between them. Every stage of the agentic engine reads from Profile Memory, and every stage can contribute new entries to it. It is how Admit Architect never forgets something you told it in week one when you come back in week eight. ## Three primitives Profile Memory has three primitives. **Facts** are statements about you. Things like "led a team of six engineers at Stripe", "scored 740 on the GMAT", "wants to work in climate investing after MBA". Each fact has a type, a value, and a confidence, and is owned by your account. **Evidence Spans** are verbatim excerpts from source documents, such as a paragraph from your resume, a sentence from an interview answer, or a line from a file you uploaded. They are stored with a pointer back to the exact location in the source. **Evidence Links** connect a fact to one or more evidence spans, so every claim in Profile Memory is auditable. You can always trace back from a fact to the exact words that support it. ## How facts get in Facts appear in Profile Memory in three ways. The engine extracts them during the Ingest stage (when you upload a resume or finish an interview). You can add one directly from a Mind session. Or Archer can propose one in the middle of another action, and the addition happens immediately because creating a new fact is an additive, safe operation. Facts are append-only in practice, but you can edit or correct one at any time. Archer will flag a fact whose evidence span no longer matches it, so stale facts are surfaced instead of silently rotting. ## How evidence links make the engine honest Because every fact is linked to evidence, every downstream stage of the engine is forced to support its claims with something concrete. Your [Dean Brief](/docs/dean-brief) cites the exact answer span behind each rubric comment. Your [Essay Workbench](/docs/essay-workbench) can show you evidence gaps before you submit a draft. Your [Connections](/docs/connections) outreach can reference specific shared context that the engine actually has evidence for. ## How it fits in the engine Profile Memory is the substrate every other stage reads from. Ingest writes to it. Interrogate writes to it. Synthesize reads and writes. Draft reads it heavily. Orchestrate reads it for matching in [Connections](/docs/connections). When you hear Archer "remembering something", this is literally where it remembers. ## Related - [Mind Sessions with Archer](/docs/mind-sessions) - [Dean Brief](/docs/dean-brief) - [Data and Privacy](/docs/data-privacy) --- # Free Tools _Source: https://admitarchitect.com/docs/free-tools_ Chances Calculator, Resume Review, Essay Grader, and Ding Analysis. No account required. Admit Architect offers four free, public tools at [admitarchitect.com/tools](https://admitarchitect.com/tools) that run on the same agentic engine as the full product. No account or payment required. They are deliberately single-shot tools: you get a high-quality evaluation for one piece of material at a time. The full agentic loop (persistent memory, multi-school orchestration, Archer sessions, proposal-apply writes) requires an account. ## Admission Chances Calculator URL: [admitarchitect.com/tools/admission-chances](https://admitarchitect.com/tools/admission-chances) Score your candidacy against 40+ MBA programs worldwide. You enter your demographics, work experience, education, test scores, goals, and your target schools. The calculator returns a per-school score, a reach/target/safety classification, a factor-by-factor breakdown, and improvement suggestions. Used by most first-time visitors as their first touchpoint. See also the full [Chances Prediction](/docs/chances-prediction) doc for the logged-in version that persists and updates. ## Resume Review URL: [admitarchitect.com/tools/resume-review](https://admitarchitect.com/tools/resume-review) Paste your resume text or upload a PDF or DOCX file. The engine returns an overall readiness rating, a per-dimension evaluation (leadership, quantified impact, career progression, communication, differentiators, gaps), and a prioritized list of specific fixes. Useful as a sanity check before you apply, or before you hand your resume to a recommender. ## Essay Grader URL: [admitarchitect.com/tools/essay-grader](https://admitarchitect.com/tools/essay-grader) Paste a school name, the essay prompt, your draft, and an optional word limit. The engine grades the draft across narrative arc, school fit, authenticity, structure, and efficiency, and returns specific suggestions. Works for any school and any prompt. ## Ding Analysis URL: [admitarchitect.com/tools/ding-analysis](https://admitarchitect.com/tools/ding-analysis) Share what happened (school, round, interview status, a short profile summary), and the engine returns a diagnosis across five categories: academic profile, professional experience, narrative and fit, application execution, and competitive positioning. You get a primary diagnosis, a re-application plan, and alternative school suggestions. ## How it fits in the engine Free Tools are single-shot entry points into the same evaluation pipeline the logged-in product uses. They write nothing to the database, persist nothing, and have no memory across runs. Their purpose is to give you a high-quality taste of the engine before you commit to an account. ## Related - [Getting Started](/docs/getting-started) - [Chances Prediction](/docs/chances-prediction) - [Pricing and Plans](/docs/pricing) --- # Data and Privacy _Source: https://admitarchitect.com/docs/data-privacy_ What Admit Architect stores, how it is encrypted, and our policy on model training. Admit Architect stores the material you upload and generate for the purpose of running the agentic engine for you. We do not sell it, we do not use it to train any model, and we encrypt it at rest and in transit. This page describes what we store, where we store it, and what you can do with it. ## What we store We store the data you give us and the data the engine derives from it. **Account.** Your email, your authentication record, and your profile settings. Authentication is handled by AWS Cognito, so your password never touches our database. **Uploads.** Any files you upload (resumes, transcripts, additional documents) are stored in a private S3 bucket. Extracted text from those files is stored in our database so the engine can reason over it. **Application state.** Every application you create, every essay draft, every version history, every recommender invitation, every deadline, every connection, and every Mind session. This is the substrate the engine operates on. **Engine outputs.** Dean Brief versions, narrative spine interviews, rubric evaluations, chances predictions, and every proposal Archer has ever made. All versioned, all auditable. ## Where it is stored Data is stored on AWS in the United States. The database runs on Amazon RDS with SSL enforced on every connection. File uploads live in a private Amazon S3 bucket with AES-256 encryption at rest. Model inference runs inside the same AWS account, so your prompts and responses never leave our environment. ## Model training We do not use your data to train any model, period. We also do not share your data with third parties for their model training. Every inference request runs in a mode that does not retain or reuse inputs for training. Archer is an inference client, not a learning system. It gets sharper at helping you specifically because of what it can read from your [Profile Memory](/docs/profile-memory), not because anyone is training a model on you. ## What you can do You can export the full state of your account (your applications, your essays, your spine, your dean brief versions) and you can delete your account and every associated record at any time from [Account Settings](/docs/account-settings). If you delete your account, extracted text, interview answers, essay drafts, facts, evidence spans, and every session are removed. Raw file uploads in S3 are also removed. A short audit trail required for abuse prevention is retained and discarded after a standard retention window. ## How it fits in the engine Privacy is not a phase of the loop, it is a constraint on every phase. Ingest only writes to Profile Memory. Synthesize only reads from memory you own. Mind sessions are scoped to your account and are never visible to any other user. If a stage of the engine ever needed to share your data for it to work, we would tell you first and let you opt out. ## Related - [Account Settings](/docs/account-settings) - [Profile Memory](/docs/profile-memory) --- # Pricing and Plans _Source: https://admitarchitect.com/docs/pricing_ How the engine's features map to each Admit Architect plan. Start free, upgrade when you want the full agentic loop. Admit Architect is self-serve SaaS with a Free tier that lets you try the engine end to end, plus paid tiers that unlock higher usage, multi-school orchestration, and the full Archer agent surface. The authoritative pricing page with current numbers is at [admitarchitect.com/pricing](https://admitarchitect.com/pricing). This doc explains how features map to tiers and when you should upgrade. ## What each tier unlocks **Free.** The full onboarding experience: account, resume upload, onboarding wizard, narrative spine interview, and enough Dean Brief generation to run the engine once end to end. The four [Free Tools](/docs/free-tools) are always free and do not count against your Free tier usage. **Pro.** The full multi-school application flow: unlimited Dean Brief regenerations, the [Essay Workbench](/docs/essay-workbench) with Archer-assisted drafting, the [Recommenders](/docs/recommenders) surface, [Deadlines and Portfolio](/docs/deadlines-portfolio), and unlimited [Mind Sessions with Archer](/docs/mind-sessions). This is the tier most applicants want. **Max.** Everything in Pro, plus the full [Connections](/docs/connections) surface, extended usage, and higher budget caps for Archer across the month. Built for candidates applying to six or more schools and people who want the engine working harder for them across the whole cycle. Exact credit counts, per-month usage caps, and dollar prices live on the [pricing page](https://admitarchitect.com/pricing) and are the source of truth. This doc will not repeat them because it should not drift. ## When to upgrade Upgrade from Free to Pro when you are done running the engine for the first time and you are ready to actually draft essays for more than one school. Before that point, the Free tier has everything you need. Upgrade from Pro to Max when you want the full Connections surface and you are committed to a multi-school portfolio where networking is a meaningful part of your strategy. ## How it fits in the engine Tiers are a ceiling on how much of the engine runs for you, not a ceiling on what the engine can do for you. Every tier runs the same loop (Ingest, Interrogate, Synthesize, Draft, Orchestrate, Submit). The higher tiers just remove constraints around multi-school, Connections, and Archer session depth. ## Related - [Free Tools](/docs/free-tools) - [Getting Started](/docs/getting-started) - [Account Settings](/docs/account-settings) --- # Account Settings _Source: https://admitarchitect.com/docs/account-settings_ Manage your profile, notifications, Archer preferences, and account deletion. Account Settings is where you manage your profile fields, your notification preferences, Archer's behavior, and your account itself. ## Profile Your profile holds the basic fields the engine uses to personalize every stage: your name, email, target round, timeline context, and post-MBA goal statement. These are the same fields you answered during onboarding, and you can edit any of them at any time. Changing a profile field does not wipe your existing application state. It simply updates the field, and the engine will use the new value the next time it reads from your profile. ## Notifications You can choose how the engine reaches out to you when something meaningful happens: a deadline is approaching, a recommender has submitted, a Dean Brief has finished regenerating, or an at-risk flag has appeared on an application. Email is the default channel. You can silence any category you do not want. Notifications are about your work, not about marketing. Product announcements are opt-in and live in a separate channel. ## Archer preferences Archer has a few knobs you can turn. You can set a default tone for its drafts (concise, conversational, analytical), you can choose how aggressive its next-action suggestions are on the dashboard, and you can restrict which write tools require your approval versus which run immediately. By default, every write except recording a new fact and completing a deadline is proposal-first. See [Mind Sessions with Archer](/docs/mind-sessions). ## Security You can change your password at any time. If you signed up with Google, your password is managed by Google and cannot be changed here. Email verification and password reset flows are handled by AWS Cognito. ## Export and delete You can export the full state of your account as a single archive, or delete your account entirely. Deletion is permanent and removes every file you uploaded, every interview answer, every essay draft, every fact, every evidence span, and every Mind session. See [Data and Privacy](/docs/data-privacy) for the full retention policy. ## How it fits in the engine Account Settings is metadata about you. It does not contain any stage of the loop, but it influences every stage. Your timeline context decides how aggressive the engine is; your Archer preferences decide how Mind sessions behave; your notification preferences decide when the engine reaches out. ## Related - [Data and Privacy](/docs/data-privacy) - [Mind Sessions with Archer](/docs/mind-sessions) - [Troubleshooting](/docs/troubleshooting) --- # Troubleshooting _Source: https://admitarchitect.com/docs/troubleshooting_ Common issues you might hit while running the engine and how to fix them. This page is a running list of the most common issues new users run into while running the agentic engine. If your problem is not here, see the [Support](#support) section at the bottom. ## Upload and extraction **My resume uploaded but no text was extracted.** Most often this is a scanned PDF, which is an image file wearing a PDF extension. Try exporting from your source (Word, Google Docs, or a resume builder) as a fresh PDF instead of scanning a printed copy. **My DOCX upload failed.** The engine supports standard DOCX files. If you have an older DOC file, save it as DOCX first. Files larger than the configured upload limit will also be rejected with a clear error. **The extracted text has strange characters.** Some PDFs embed fonts in ways that mangle copy-paste extraction. If you see garbled text, the easiest fix is to open the file in a word processor, save a fresh copy, and re-upload. ## Narrative spine interview **The interview is asking follow-up questions I do not want to answer.** The engine pushes back when an answer is surface-level. If you are sure you want to move on, you can end the current question with a clear answer and the engine will accept it. Depth is encouraged, not enforced. **My answer got cut off or is missing.** The interview is append-only. Reopen the same question and the previous version should still be there as a saved answer. If it is truly missing, open a [Mind session](/docs/mind-sessions) and ask Archer to list your interview answers. ## Dean Brief **Generation is stuck on the same stage.** The Dean Brief runs a staged pipeline and each stage has its own status. If it has been stuck longer than a minute or two, refresh the page. Versions are append-only, so nothing is lost if you need to retry. **The brief cites evidence I do not recognize.** Open the evidence citation directly. Every claim links back to a specific span in [Profile Memory](/docs/profile-memory). If the span is wrong, you can correct the underlying fact and regenerate. ## Essay Workbench **I lost a change Archer made.** Every save creates a new version. Open the version history for the submission and restore the draft you wanted. **Archer's proposal is not appearing.** Proposals arrive as a stream. If your connection drops mid-response, the stream can be cut. Try the request again. If it still fails, check [Account Settings](/docs/account-settings) to confirm you are under your current plan's usage cap. ## Recommenders **A recommender did not receive their invite.** Ask them to check spam first. You can also copy the invite link directly from the [Recommenders](/docs/recommenders) surface and send it through your own channel. **A recommender opened the link but cannot see the questions.** Make sure you have selected a target school for the application they are recommending for. Without a school selected, the engine has no template to show them. ## Account and billing **I cannot sign in.** Password reset is handled by AWS Cognito through the sign-in page. If you signed up with Google, sign in with Google, not a password. **I was charged for a plan I did not mean to select.** Write to support from the email on your account within the refund window on your [plan](https://admitarchitect.com/pricing) and we will sort it out. ## Support If nothing here resolves your issue, reach out to support at [admitarchitect.com](https://admitarchitect.com). Include your account email, the page you were on, and the specific thing you were trying to do. The more specific the report, the faster we can fix the underlying issue in the engine. ## Related - [Account Settings](/docs/account-settings) - [Data and Privacy](/docs/data-privacy) - [Mind Sessions with Archer](/docs/mind-sessions)