Skip to content

Configuration Guide

Jump to bottom
AbhishekMandapmalvi edited this page Mar 11, 2026 · 1 revision

Configuration Guide

Config File Location

All configuration is stored in a single JSON file:

~/.autoapply/config.json

On Windows this resolves to C:\Users\<username>\.autoapply\config.json.

The data directory can be overridden with the AUTOAPPLY_DATA_DIR environment variable.

Tip: You can edit config.json directly, but it is recommended to use the Settings UI or API to avoid validation errors. The app validates configuration using Pydantic v2 models on load.

LLM Provider Setup

AutoApply supports four LLM providers for AI-powered resume and cover letter generation.

Supported Providers

Provider Default Model API Key Prefix
Anthropic claude-sonnet-4-20250514 sk-ant-
OpenAI gpt-4o sk-
Google gemini-2.0-flash AI...
DeepSeek deepseek-chat sk-

Configuration

Set the provider in the llm section of config.json:

{
  "llm": {
    "provider": "anthropic",
    "api_key": "sk-ant-...",
    "model": "claude-sonnet-4-20250514"
  }
}

Or configure via the Settings page in the UI (Settings > AI Provider).

You can change the model to any model supported by the provider. For example:

{
  "llm": {
    "provider": "openai",
    "api_key": "sk-...",
    "model": "gpt-4o-mini"
  }
}

Validate Your Key

Use the API to validate before saving:

curl -X POST http://127.0.0.1:5000/api/validate-api-key \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{"provider": "anthropic", "api_key": "sk-ant-..."}'

No AI Provider

If no provider is configured, AutoApply falls back to template-based document generation. Resumes and cover letters use your profile data directly without AI tailoring.

API Key Storage

AutoApply stores API keys securely using the OS keyring when available:

Platform Backend
Windows Windows Credential Locker
macOS macOS Keychain
Linux SecretService (GNOME Keyring / KWallet)

If the keyring is unavailable (e.g., headless Linux without a desktop environment), keys fall back to plaintext storage in config.json.

Auto-migration: On startup, if a plaintext API key is found in config.json and the keyring is available, the key is automatically migrated to the keyring and removed from the config file.

User Profile

The profile section stores your personal information used for applications:

{
  "profile": {
    "full_name": "Jane Doe",
    "email": "jane@example.com",
    "phone": "+1-555-0100",
    "location": "San Francisco, CA",
    "linkedin_url": "https://linkedin.com/in/janedoe",
    "years_experience": 5,
    "education": "BS Computer Science, Stanford University",
    "skills": ["Python", "JavaScript", "TypeScript", "AWS", "Docker"],
    "screening_answers": {
      "Are you authorized to work in the US?": "Yes",
      "Do you require visa sponsorship?": "No",
      "What is your expected salary?": "$150,000",
      "Are you willing to relocate?": "Yes",
      "years_of_experience_python": "5",
      "years_of_experience_javascript": "4"
    }
  }
}

Field Reference

Field Type Required Description
full_name string Yes Your full name as it should appear on applications
email string Yes Email address for applications
phone string No Phone number
location string Yes City, State/Country
linkedin_url string No Your LinkedIn profile URL
years_experience int Yes Total years of professional experience
education string No Education summary
skills string[] No List of skills
screening_answers object No Pre-filled answers for common screening questions

Screening Answers

The screening_answers dictionary is used by Workday and Ashby appliers to auto-fill screening questions during the application process. Keys are the question text (or a normalized form), and values are your answers.

You can add screening answers via the Settings UI (Profile > Screening Answers) or directly in config.json.

Search Criteria

The search_criteria section defines what jobs the bot looks for:

{
  "search_criteria": {
    "job_titles": ["Software Engineer", "Backend Developer", "Full Stack Developer"],
    "locations": ["San Francisco, CA", "New York, NY", "Remote"],
    "remote_only": false,
    "experience_level": "mid",
    "excluded_companies": ["Company A", "Company B"]
  }
}

Field Reference

Field Type Default Description
job_titles string[] [] Job titles to search for
locations string[] [] Geographic locations to search in
remote_only bool false Only return remote positions
experience_level string "mid" Filter: entry, mid, senior, executive
excluded_companies string[] [] Companies to skip

Bot Settings

The bot_settings section controls bot behavior:

{
  "bot_settings": {
    "apply_mode": "review",
    "max_applications_per_day": 50,
    "search_engines": ["linkedin", "indeed"]
  }
}

Apply Modes

Mode Behavior
full_auto Searches, scores, generates documents, and applies automatically. No human intervention needed.
review Searches and scores jobs, then pauses for your approval. You approve, reject, skip, or manually apply for each job.
watch Searches and scores only. No applications are submitted. Useful for monitoring the job market.

Field Reference

Field Type Default Description
apply_mode string "review" One of: full_auto, review, watch
max_applications_per_day int 50 Daily application cap
search_engines string[] ["linkedin", "indeed"] Platforms to search

Schedule Configuration

The schedule section enables time-based automatic bot operation:

{
  "schedule": {
    "enabled": true,
    "start_time": "09:00",
    "end_time": "17:00",
    "days_of_week": [0, 1, 2, 3, 4],
    "timezone": "America/Los_Angeles"
  }
}

Field Reference

Field Type Default Description
enabled bool false Enable scheduled operation
start_time string "09:00" Time to auto-start (HH:MM, 24-hour)
end_time string "17:00" Time to auto-stop (HH:MM, 24-hour)
days_of_week int[] [0,1,2,3,4] Days to run (0=Monday, 6=Sunday)
timezone string "UTC" IANA timezone identifier

When scheduling is enabled, the bot automatically starts at start_time and stops at end_time on the configured days. Outside the schedule window, the bot remains idle.

Environment Variables

Variable Default Description
AUTOAPPLY_DEV unset Set to 1 to enable development mode. Bypasses API authentication, enables debug logging.
AUTOAPPLY_DEBUG unset Set to 1 to enable Flask debug mode and verbose logging.
AUTOAPPLY_LOG_FORMAT text Log output format: text (human-readable) or json (structured).
AUTOAPPLY_DATA_DIR ~/.autoapply Override the data directory path. Affects config, database, browser profile, and experience files locations.

Example Usage

# Run in development mode with JSON logging
export AUTOAPPLY_DEV=1
export AUTOAPPLY_LOG_FORMAT=json
python run.py

# Use a custom data directory
export AUTOAPPLY_DATA_DIR=/opt/autoapply/data
python run.py

Full Config Example

{
  "profile": {
    "full_name": "Jane Doe",
    "email": "jane@example.com",
    "phone": "+1-555-0100",
    "location": "San Francisco, CA",
    "linkedin_url": "https://linkedin.com/in/janedoe",
    "years_experience": 5,
    "education": "BS Computer Science, Stanford University",
    "skills": ["Python", "JavaScript", "AWS"],
    "screening_answers": {
      "Are you authorized to work in the US?": "Yes",
      "Do you require visa sponsorship?": "No"
    }
  },
  "search_criteria": {
    "job_titles": ["Software Engineer", "Backend Developer"],
    "locations": ["San Francisco, CA", "Remote"],
    "remote_only": false,
    "experience_level": "mid",
    "excluded_companies": []
  },
  "bot_settings": {
    "apply_mode": "review",
    "max_applications_per_day": 50,
    "search_engines": ["linkedin", "indeed"]
  },
  "llm": {
    "provider": "anthropic",
    "model": "claude-sonnet-4-20250514"
  },
  "schedule": {
    "enabled": false,
    "start_time": "09:00",
    "end_time": "17:00",
    "days_of_week": [0, 1, 2, 3, 4],
    "timezone": "America/Los_Angeles"
  }
}

AutoApply Wiki

User Guide

Technical

Build & Deploy

Project

Clone this wiki locally