Developers

Resume Tailor, as an API and an MCP server.

The résumé tailor from the Promptly app, open to your scripts, your apps and AI agents. Send a résumé and a job description; get every rewrite next to the original, anything it had to invent flagged, and a PDF or .tex to download.

Any plan, Free included · 1 résumé credit per tailoring · revising, editing and building are free for 24 hours

Authentication

Every request sends your Promptly license key as a Bearer token: Authorization: Bearer plk_…. Find it on your account page. It's the same key the desktop app uses, so keep it private.

REST quick start

KEY=plk_...   # your license key, from trypromptly.app/account
# One call: résumé file + job description in, tailored PDF link out
curl -s https://trypromptly.app/v1/tailor/tailorings \
  -H "Authorization: Bearer $KEY" -H "content-type: application/json" -d @- <<EOF
{"resume": {"file": {"name": "resume.pdf", "base64": "$(base64 -w0 resume.pdf)"}},
 "job_title": "Staff Backend Engineer",
 "job_description": $(jq -Rs . < job.txt),
 "build": "pdf"}
EOF

The answer lists every change before and after, with invented text flagged, plus download links. Clients with short timeouts (GPT Actions, for one) can send "wait": false, get a 202 at once, and poll GET /v1/tailor/tailorings/{id}?wait=30. The full schema is in the OpenAPI spec, ready to import into Postman, GPT Actions or an agent framework.

POST /v1/tailor/resumesImport a résumé (text, LaTeX, or a PDF/Word file as base64) → an id and its rewritable items. Free, 20 a day.
POST /v1/tailor/tailoringsTailor to a job: resume_id or an inline resume, job_title, job_description, optional sections, wait, build. 1 credit.
GET /v1/tailor/tailorings/{id}?wait=30Read a tailoring; wait long-polls one that's still running.
POST …/{id}/revise · /edit · /shortenRewrite one item again, set exact wording or keep originals, or shorten the accepted rewrites.
POST …/{id}/buildBuild a PDF or .tex → download links (no auth needed, they expire with the tailoring).
GET /v1/tailor/creditsCredits left.
DELETE /v1/tailor/resumes/{id}Delete a résumé and everything built from it, now.

MCP server

Streamable HTTP at https://trypromptly.app/mcp, with the same operations as tools: tailor_resume, import_resume, revise_resume_item, edit_resume_items, shorten_resume, build_resume, get_tailoring, get_resume_credits and delete_resume. A tailoring takes 30 to 90 seconds; clients that accept event streams get progress notifications while it runs.

claude mcp add --transport http promptly-resume https://trypromptly.app/mcp --header "Authorization: Bearer plk_..."
// Cursor: .cursor/mcp.json (VS Code: .vscode/mcp.json, under "servers" with "type": "http")
{ "mcpServers": { "promptly-resume": {
    "url": "https://trypromptly.app/mcp",
    "headers": { "Authorization": "Bearer plk_..." } } } }
// Claude Desktop (stdio config) through the mcp-remote bridge
{ "mcpServers": { "promptly-resume": {
    "command": "npx",
    "args": ["-y", "mcp-remote", "https://trypromptly.app/mcp", "--header", "Authorization:${PROMPTLY_AUTH}"],
    "env": { "PROMPTLY_AUTH": "Bearer plk_..." } } } }
// Claude API (MCP connector) or OpenAI Responses API
"mcp_servers": [{ "type": "url", "url": "https://trypromptly.app/mcp",
  "name": "promptly-resume", "authorization_token": "plk_..." }]

"tools": [{ "type": "mcp", "server_label": "promptly-resume",
  "server_url": "https://trypromptly.app/mcp",
  "headers": { "Authorization": "Bearer plk_..." } }]

What's stored, and for how long

Résumés, tailorings and the files built from them are kept for 24 hours after their last use so follow-up calls can revise and download them, then deleted automatically. They're never copied into backups, and DELETE /v1/tailor/resumes/{id} removes them sooner. Download links are unguessable, need no auth so an agent can hand them to a person, and stop working when the tailoring is deleted. See the privacy policy.

Questions

Does it make things up?

It's told not to, and anything it couldn't take from your résumé (a number, a tool, a scope) comes back listed in invented on that change, so you or your agent can check it before keeping it.

Which résumé formats work?

PDF, Word (.docx), LaTeX and plain text. A LaTeX résumé keeps its own template: rewrites are spliced into your source. Everything else is rebuilt in a clean one-column template.

What does it cost?

One résumé credit per tailoring, from the same balance as the app: Free includes 3, Base 20 a month and Pro 60, and credit packs top up any plan. Importing, revising, editing, shortening and building are free for 24 hours after a tailoring.

Can I use it from claude.ai or ChatGPT's connector directory?

Not yet. Those connectors sign in with OAuth, and this server accepts license keys only. Claude Code, Cursor, VS Code, Claude Desktop (through mcp-remote) and the Claude and OpenAI APIs all work today.

Prefer clicking? The same tailor is in the desktop app: Resume Tailor.

You're sharing your screen. Promptly isn't in it.