Podman-First FastAPI Service

Resume Ops Documentation

Headless API service for tailoring JSON Resumes using a structured LangGraph state machine while strictly protecting immutable resume fields.

🎯 Signal Over Noise

Re-aligns your authentic experience to match the target job description so your true fit is immediately clear to recruiters—without keyword stuffing or spam.

🛡️ Truth Over Fabrication

Your contact details, job titles, companies, dates, and degrees are strictly immutable. resume-ops reformulates real achievements; it never invents fake projects or unverified skills.

🔒 Privacy & Self-Hosting

Runs in your own isolated Podman/Docker container. Retain full control over model routing—whether running 100% offline with local LLMs (vLLM, Ollama) or routing to your trusted API provider of choice.

Deployment Snippets

Copy-pasteable setup commands for Podman, Docker Compose, systemd Quadlets, and API requests.

podman run -d \
  --name resume-ops \
  -p 8000:8000 \
  -v ./data:/data:Z,U \
  --env-file .env \
  ghcr.io/rat-s/resume-ops:latest

# Run pre-built GHCR registry image via compose
podman compose up -d

# Service endpoints:
# - API: http://localhost:8000
# - Health check: http://localhost:8000/healthz

# ~/.config/containers/systemd/resume-ops.container
[Unit]
Description=Resume Ops FastAPI Container
After=network-online.target

[Container]
Image=ghcr.io/rat-s/resume-ops:latest
PublishPort=8000:8000
Volume=%h/resume-ops/data:/data:Z,U
EnvironmentFile=%h/resume-ops/.env
AutoUpdate=registry

[Install]
WantedBy=default.target

curl -X POST http://localhost:8000/api/v1/tailor \
  -H "Content-Type: application/json" \
  -d '{
    "job_description": "Senior DevOps Engineer with Kubernetes & Terraform experience",
    "theme": "jsonresume-theme-folio"
  }'

curl -X POST http://localhost:8000/api/v1/tailor \
  -H "Content-Type: application/json" \
  -d '{
    "job_description": "Full Stack Lead Engineer",
    "callback_url": "http://job-ops:3005/api/webhooks/resume-ready"
  }'

Interactive .env Configurator

Generate a complete .env configuration file matching .env.example with live model fetching from OpenRouter or custom OpenAI base URLs.

1. API Keys & Base URL

OPENAI_BASE_URL

Queries /v1/models to auto-populate available LLM models into dropdowns.

OPENAI_API_KEY / OpenRouter Key

ANTHROPIC_API_KEY

GEMINI_API_KEY

2. Model Routing

DEFAULT_MODEL

STRATEGY_MODEL (Optional Override)

WORK_MODEL (Optional Override)

SKILLS_MODEL (Optional Override)

3. Theme, Tone & Rate Limits

DEFAULT_THEME

TAILORING_STYLE

LLM_MAX_CONCURRENCY

LLM_MAX_RETRIES

Generated .env File

Tailoring & Field Protection Rules

Conservative boundaries prevent LLM hallucinations while aligning resume content to job requirements.

JSON Resume Section	Status	Tailoring Behavior
`basics`	🔒 Protected	Name, email, phone, and links are 100% preserved. Only `summary` is tailored.
`work`	⚡ Tailored	Preserves all job titles, dates, and companies. Tailors `summary` and `highlights`.
`education`	⚡ Tailored	Degrees, dates, and institutions remain immutable. Only relevant `courses` are tailored.
`skills`	⚡ Tailored	Regrouped and prioritized based on target JD, strictly derived from master resume.
`projects`	⚡ Tailored	Selected and reordered. New unverified projects are never hallucinated.
`certificates`	🔒 Selection Only	Content is not rewritten; final output is capped at maximum 18 items.

API Endpoints

FastAPI OpenAPI endpoints available on running instances.

POST /api/v1/tailor

Initiates resume tailoring in synchronous or asynchronous webhook mode.

GET /api/v1/tasks/{task_id}

Polls execution status and fetches completed output of background tasks.

GET /api/v1/themes

Lists installed and data-mounted JSON Resume themes available for PDF rendering.