create_resume
ChatGPTCreate a new, editable resume for the user and return it ready to build on. Starts a fresh resume with the Personal Information header visible (seeded with any details you pass) and every other standard section — Summary, Work Experience, Education, Skills, and so on — hidden. Each hidden section reveals itself the first time you fill it with update_section, so after creating you add content section by section. Returns the new resume's id (pass it to the other tools) plus a preview. Args: name: A title for the resume, e.g. "Backend Engineer - Acme". template_id: Layout template id (see list_templates). Defaults to 14. first_name: Optional - seeds the header first name. last_name: Optional - seeds the header last name. headline: Optional - seeds the professional headline shown under the name. email: Optional - seeds the contact email. Returns: The new resume: its resume_id, name, and a rendered preview.
delete_entry
ChatGPTDelete one entry from a resume section (e.g. an old job or a duplicate). Later entries shift down to fill the gap. Deleting a section's last entry hides the section. This is the only way to REMOVE an entry — blanking its fields with update_section leaves an empty record behind. Args: resume_id: UUID of the resume. section_code: e.g. "work-experience", "education", "summary". index: Which entry to delete (0-based, as shown by get_section). Returns: The updated resume, as a live preview + text summary.
export_resume_pdf
ChatGPTExport a resume as a PDF and return a download link. The link is short-lived. In a UI-capable host the preview's Download button opens it for the user; otherwise the link is in the text response so the user (or the model) can present it. Args: resume_id: UUID of the resume to export. Returns: JSON with the download URL and filename (also surfaced as text).
get_job
ChatGPTOne job in full: the description, requirements, pay, location, and whether it fits the user's preferences. Args: job_id: The job_id from search_jobs. Returns: The job's details and its apply link.
get_resume
ChatGPTRead a full resume: its sections, entries, and styling. Renders the styled resume inline on hosts that support it; on other hosts returns a readable plain-text rendering of the whole document. Args: resume_id: UUID of the resume. Returns: The resume, as a live preview + text summary.
get_section
ChatGPTRead one section of a resume. Args: resume_id: UUID of the resume. section_code: One of: personal-information, summary, work-experience, education, skill, certification, award, achievements, project, volunteering, publications, references. Returns: JSON with that section's entries, and the field names it accepts.
list_job_filters
ChatGPTThe values search_jobs accepts, and an optional lookup of places the job board has jobs in. Args: location: Optional start of a place name, e.g. "beng" or "germ", to see matching locations. Returns: Accepted values for remote, seniority, employment_type and sort, plus matching locations if asked.
list_resumes
ChatGPTList the user's resumes, most recently updated first. Args: limit: How many to return (default 20, max 50). Returns: JSON with each resume's id, name, template, and last-updated date.
list_templates
ChatGPTList the resume templates available to this user. Returns: JSON with each template's id, name, code, and whether it needs a paid plan.
reorder_entries
ChatGPTRearrange the entries WITHIN one section, in a single call. Give the complete new order as the current 0-based entry positions (as shown by get_section). Examples for a 4-entry section: move the last entry to the top -> [3, 0, 1, 2] swap the first two -> [1, 0, 2, 3] Args: resume_id: UUID of the resume. section_code: A repeatable section — e.g. "project", "work-experience", "education", "skill", "certification". order: Every current entry position exactly once, in the new sequence. Returns: The updated resume, as a live preview + text summary.
reorder_sections
ChatGPTSet the order sections appear in on the resume. Args: resume_id: UUID of the resume. section_codes: All visible section codes, in the order you want them. Any section you leave out becomes hidden. Returns: JSON with the new order.
rewrite_section
ChatGPTRewrite one resume entry's description with AI, and save it. Uses ResumeUp's own writing models, so the result matches what the web app produces. Costs 1 credit unless the user is subscribed. Args: resume_id: UUID of the resume. section_code: Which section to rewrite — e.g. "summary", "work-experience", "project". index: Which entry within that section (0-based). Default 0. instructions: Optional steer, e.g. "make it more concise" or "emphasise leadership". job_description: Optional job posting to angle the writing towards. Returns: JSON with the rewritten content and a fresh plain-text rendering.
search_jobs
ChatGPTSearch the job board for jobs that fit the user, best matches first. Anything not given falls back to the user's saved job preferences (roles, locations, work style), so a call with no arguments returns jobs matching their profile. Results carry no scores; the order is the ranking. Args: job_titles: Roles to search for, e.g. ["Backend Engineer", "Data Analyst"] (up to 5). keywords: Free-text words that must appear, e.g. "python kubernetes". locations: Places, e.g. ["Bengaluru", "Remote US", "Germany"] (up to 5). remote: Any of "onsite", "hybrid", "remote". seniority: Any of intern, entry, mid, senior, staff, principal, director, executive. employment_type: Any of full_time, part_time, contract, internship, temporary. posted_within_days: Only jobs posted in the last N days (1-90). sort: "match" (best fit first, default) or "newest". limit: How many jobs to return (1-10, default 10). Returns: The matching jobs with their job_id (use get_job for details, apply_to_job to apply).
set_template
ChatGPTSwitch a resume to a different template. Re-renders the resume in the new template on hosts that support the preview. Args: resume_id: UUID of the resume. template_id: Numeric template id from list_templates (NOT the code). Returns: The resume in its new template, as a live preview + text summary.
tailor_to_job
ChatGPTTailor a resume to a specific job description. Creates a NEW resume tailored to the job and leaves the original untouched, so nothing the user already has is overwritten. Costs 3 credits unless the user is subscribed. Takes up to a few minutes. Args: resume_id: UUID of the resume to tailor FROM. job_description: The full job posting text to tailor towards. new_name: Name for the tailored copy. Defaults to "<original name> (tailored)". Returns: JSON with the new resume's id and its tailored content.
twinny_add_fact
ChatGPTUse this when the person wants to teach their Twinny something new about themselves. Their Twinny may tell anyone what it's taught. Save only what they actually said; never add details they didn't give.
twinny_add_source
ChatGPTUse this when the person wants their Twinny to learn from a web page or a piece of text, e.g. their portfolio, a project write-up or notes. Their Twinny may answer anyone from it. Give exactly one of url or text. Reading usually takes under a minute.
twinny_answer_question
ChatGPTUse this when the person answers a question people asked their Twinny. Their Twinny learns the answer and the question leaves their inbox. Save only what they actually said.
twinny_ask
ChatGPTUse this when someone wants to ask a person's Twinny a question, e.g. a recruiter checking a candidate's experience. The Twinny answers in first person, only from what that person made public, and names what it used. A question it can't answer goes to the person's inbox so they can add the answer. Pass conversation_id from the previous answer for follow-ups. Takes 10–20 seconds.
twinny_compare_to_job
ChatGPTUse this when someone wants to see how one person's Twinny fits a job. Reads the job into must-have and nice-to-have requirements, checks each against what the person made public, and quotes the evidence. Returns a 0–100 score, a match tier and the requirement checklist. 'No public evidence' means their Twinny doesn't say, not that they lack it. A person should review it before any hiring decision. Takes 15–30 seconds.
twinny_create
ChatGPTUse this when the person wants a Twinny and doesn't have one yet. Builds it from their ResumeUp profile in about 30 seconds. It stays private until they publish it. If they already have one, it is returned unchanged.
twinny_dismiss_question
ChatGPTUse this when the person wants to clear a question from their inbox without answering it.
twinny_edit_fact
ChatGPTUse this when the person wants to correct one thing their Twinny knows. To keep something from their Twinny, remove it instead.
twinny_get_my_twin
ChatGPTUse this when the person asks about their own Twinny (their AI twin on ResumeUp): whether they have one, whether it's live, its link, how it introduces them, its settings, and how many questions are waiting. Call it before publishing so you can show them what will be public.
twinny_list_knowledge
ChatGPTUse this when the person wants to see or change what their Twinny knows. Lists every fact (with its id and whether it's public or private) and every document, note or link it learned from. Their Twinny may tell anyone what's public; private facts are sensitive ones (health, family, other people) it marks itself and never shares. Use the ids with the edit and remove tools.
twinny_list_questions
ChatGPTUse this when the person wants to see the questions people asked their Twinny that it couldn't answer. Answering them (twinny_answer_question) teaches their Twinny so it can answer next time.
twinny_publish
ChatGPTUse this only when the person clearly asks to publish their Twinny, or to republish after changes. It makes their Twinny public at its link: anyone with the link can chat with it, and it answers only from what they marked public. Before calling, show them what will be public (twinny_get_my_twin, twinny_list_knowledge) and get a clear yes. Takes up to 45 seconds.
twinny_rank_for_job
ChatGPTUse this when someone wants to compare several people's Twinnys for one job. Each Twinny is checked on its own against the same requirements, from what that person made public, then they're ordered by score. Returns each one's score, tier and requirement verdicts. A person should review it before any hiring decision. Takes 20–45 seconds.
twinny_remove_fact
ChatGPTUse this when the person wants their Twinny to forget one fact. Confirm which fact first.
twinny_remove_source
ChatGPTUse this when the person wants their Twinny to forget a document, note or link it learned from. Confirm which one first.
twinny_unpublish
ChatGPTUse this when the person asks to take their Twinny offline. The link stops working; nothing it knows is deleted, and they can publish again any time.
twinny_update_settings
ChatGPTUse this when the person wants to change how their Twinny talks (tone, answer length), whether it shares their profile links, or what it says about pay, notice period, work permit and relocation. Only the settings given change.
update_section
ChatGPTCreate or update one entry in a resume section. Only the fields you pass are changed; everything else in the entry is left alone. Descriptions must be HTML — use <ul><li>…</li></ul> for bullets or <p>…</p> for prose, because that is how the resume templates render them. To REARRANGE entries, do not rewrite them one by one — use reorder_entries (one call). To REMOVE an entry, use delete_entry. Args: resume_id: UUID of the resume. section_code: e.g. "work-experience", "summary", "skill". fields: Field-name -> value, e.g. {"job-title": "Staff Engineer", "description": "<ul><li>…</li></ul>"}. Call get_section first to see the field names a section accepts. For "skill", "values" takes a list of strings. index: Which entry to edit (0-based). For repeatable sections (work-experience, education, skill, ...), omit to APPEND a new entry. Single-entry sections (summary, personal-information, achievements) always edit their one entry — index is not needed. Returns: JSON with the updated section and a fresh plain-text rendering.