applika-cli 0.1.2

Job application tracker CLI for Applika.dev

applika-cli

Command-line interface for Applika.dev — a structured job application tracker designed for active job seekers who want full control over their data without relying on a browser.

Track every application you send, filter and review your pipeline from the terminal, and keep your AI coding assistant (Claude Code, Gemini, Codex) aware of the CLI through a bundled skill file.

Requirements

• Python 3.10+
• uv (recommended) or pip

Installation

# Using uv (recommended)
uv tool install applika-cli

# Using pipx
pipx install applika-cli

This installs the applika binary globally. Verify:

applika --help

To install from a local source checkout (useful during development):

uv tool install --force .

---

Authentication

The CLI authenticates via GitHub OAuth. On applika login, your browser opens to GitHub's authorization page. Once you authorize, the callback is captured locally on a loopback port and the session (access + refresh tokens) is saved to ~/.config/applika/session.json. The session refreshes automatically on expiry — you only need to log in once per device.

# Open browser and complete GitHub OAuth
applika login

# Verify the active session without making any other API call
applika whoami
# → Logged in as: username=luissoares  name=Luis Soares  email=[e-mail]

# Revoke the session server-side and clear local storage
applika logout

applika whoami is the recommended pre-flight check. Run it before any other command, especially in scripts or AI-assisted workflows, to confirm a valid session exists before hitting the API.

---

Commands

applika applications list

Lists all job applications in the current cycle, sorted by date descending. Supports rich filtering and both human-readable table output and machine-readable JSON.

# Default: all applications as a table
applika applications list

# Narrow down by any combination of filters
applika applications list \
  --search "stripe" \
  --mode active \
  --status active \
  --platform LinkedIn \
  --from 2026-01-01 \
  --to 2026-06-30

# JSON output — useful for piping into jq or scripts
applika applications list --output-format json

# Scope to a specific job-search cycle by UUID
applika applications list --cycle-id <uuid>

Filter reference:

Flag	Values	Default	Description
--search TEXT	any string	—	Case-insensitive substring match on company name or role title
--mode	active · passive · all	all	active = you applied; passive = recruiter reached out
--status	active · finalized · all	all	finalized applications have a recorded outcome
--platform TEXT	e.g. LinkedIn	—	Exact match on platform name
--from YYYY-MM-DD	date	—	Include applications from this date (inclusive)
--to YYYY-MM-DD	date	—	Include applications up to this date (inclusive)
--output-format	table · json	table	json returns the raw API response as a formatted array
--cycle-id TEXT	UUID	—	Filter to a specific job-search cycle

---

applika applications new

Records a new job application. The CLI validates the payload locally with Pydantic before calling the API, so you get clear field-level error messages without a round-trip.

# Minimal — required fields only
applika applications new \
  --company "Stripe" \
  --role "Backend Engineer" \
  --platform "LinkedIn" \
  --mode active \
  --date 2026-05-10

# With salary info (currency and period are always required together with any salary field)
applika applications new \
  --company "Cloudflare" \
  --role "Systems Engineer" \
  --platform "Email" \
  --mode passive \
  --date 2026-05-10 \
  --observation "Recruiter cold-messaged. Interesting stack." \
  --country "Brazil" \
  --work-mode remote \
  --salary-min 18000 \
  --salary-max 24000 \
  --currency BRL \
  --salary-period monthly

Required flags:

Flag	Description
--company TEXT	Company name. Matched against known companies; a new record is created if not found.
--role TEXT	Job title or role description
--platform TEXT	Platform where you found or were contacted about the role (e.g. LinkedIn, Indeed, Email)
--mode	active — you applied proactively · passive — inbound from a recruiter
--date YYYY-MM-DD	Date the application was submitted or the first contact occurred

Optional flags:

Flag	Description
--company-url URL	Company website
--job-url URL	Direct link to the job posting
--observation TEXT	Free-form notes about the role, process, or company
--country TEXT	Country where the role is based
--work-mode	remote · hybrid · on_site
--experience-level	intern · junior · mid_level · senior · staff · lead · principal · specialist
--expected-salary FLOAT	The salary you expect or asked for
--salary-min FLOAT	Lower bound of a posted salary range
--salary-max FLOAT	Upper bound of a posted salary range
--currency	USD · BRL · EUR · GBP · CAD · AUD · JPY · CHF · INR
--salary-period	hourly · monthly · annual

Salary rule: if any salary amount is provided (--expected-salary, --salary-min, or --salary-max), both --currency and --salary-period become required. The CLI enforces this before making any API call.

---

applika applications edit <id>

Updates an existing application. Only the flags you pass are changed — everything else keeps its current value. This makes partial updates safe: you can update just the role name or add salary info without touching anything else.

# Update the role title
applika applications edit 42 --role "Staff Engineer"

# Add salary info that wasn't captured at application time
applika applications edit 42 \
  --expected-salary 180000 \
  --currency USD \
  --salary-period annual

# Clear fields you no longer want to track
applika applications edit 42 \
  --clear-job-url \
  --clear-observation \
  --clear-country \
  --clear-salary

Find the application id with:

applika applications list --search "company name" --output-format json

Clear flags — set a field back to null without affecting others:

Flag	Clears
--clear-job-url	Job posting URL
--clear-observation	Notes
--clear-country	Country
--clear-salary	All salary fields (expected_salary, salary_min, salary_max, currency, salary_period)

Finalized applications (those with a recorded outcome) are read-only and cannot be edited. The CLI checks this before sending the request.

---

applika skill

Installs the bundled AI skill into your assistant's skills directory. The skill teaches Claude Code, Gemini, or Codex how to use this CLI — what commands exist, how authentication works, required vs optional flags, and common workflows.

The skill file is shipped inside the installed package (applika/skills/applika-cli/SKILL.md) so it stays in sync with the CLI version you have installed. By default the command creates a symlink so updates are reflected automatically; it falls back to a file copy if symlink creation fails (e.g. Windows without Developer Mode enabled).

# Interactive — choose which tool(s) to install for
applika skill
# →  1. Claude   (~/.claude/skills/applika-cli)
# →  2. Gemini   (~/.gemini/skills/applika-cli)
# →  3. Codex    (~/.codex/skills/applika-cli)
# →  4. All of the above

# Install to the current project's .claude/skills/ (file copy, no prompt)
# Useful when you want the skill scoped to a single repo
applika skill --local

# Install to any arbitrary directory (file copy, no prompt)
applika skill --dir /path/to/skills

# Preview what would be installed without touching the filesystem
applika skill --dry-run

# Overwrite an existing installation
applika skill --force

Once installed, the AI assistant automatically loads the skill context in every session and knows how to:

• check authentication with applika whoami before running commands
• construct valid new and edit payloads
• apply the correct filters on list
• recover from auth errors by prompting for applika login

---

Global options

These flags apply to every command and are passed before the subcommand name:

applika --api-base-url https://staging.applika.dev/api applications list

Flag	Default	Env variable
--api-base-url TEXT	https://applika.dev/api	APPLIKA_API_BASE_URL

---

Package structure

cli/
├── pyproject.toml               # Build config, dependencies, entry point
├── Makefile                     # Shortcuts: test, lint, format
├── README.md
├── CLAUDE.md                    # Guidance for AI assistants working in this repo
└── src/
    └── applika/                 # Importable package (entry: applika.main:main)
        ├── main.py              # Entry point — calls app()
        ├── app.py               # Root Typer app, --api-base-url callback, AppConfig wiring
        ├── config.py            # AppConfig dataclass + resolve_api_base_url()
        │
        ├── skills/
        │   └── applika-cli/
        │       └── SKILL.md     # Bundled AI skill — installed via `applika skill`
        │
        ├── schemas/             # Vendored Pydantic models (no backend import)
        │   ├── enums.py         # StrEnum types: Currency, SalaryPeriod, WorkMode, etc.
        │   └── application.py   # ApplicationCreate, ApplicationUpdate with validators
        │
        ├── lib/                 # Infrastructure — no Typer dependency
        │   ├── api.py           # ApiClient (httpx + cookie auth), ApiError, AuthError
        │   ├── session.py       # SessionData, SessionStore (~/.config/applika/session.json)
        │   └── loopback.py      # LoopbackLoginServer for OAuth browser callback
        │
        ├── utils/               # Pure helpers — no httpx, no Typer
        │   ├── dates.py         # parse_date, ensure_date_string
        │   └── output.py        # render_application_table, print_application_summary
        │
        └── commands/
            ├── auth.py          # login, logout, whoami commands
            ├── skill.py         # skill command — installs the AI skill
            └── applications/
                ├── __init__.py  # applications_app Typer sub-app, default-to-list callback
                ├── commands.py  # list_applications, new_application, edit_application
                ├── filter.py    # filter_applications, resolve_platform_id
                └── payloads.py  # ApplicationArgs dataclass, build_application_payload()

Layer rules:

• lib/ — HTTP and session logic only. No Typer, no output formatting.
• utils/ — Pure functions. No side effects, no I/O beyond what the function signature implies.
• schemas/ — Vendored copies of backend DTOs. Keep in sync manually with backend/app/application/dto/application.py and backend/app/core/enums.py when the backend changes.
• commands/ — Typer command functions only. Delegates to lib/ and utils/.

---

Development

# Install project + dev dependencies into a local virtualenv
uv sync

# Run the test suite
make test

# Auto-fix lint issues
make lint

# Apply code formatter
make format

# Build a wheel for distribution
uv build

# Install the locally built CLI globally for manual testing
uv tool install --force .

Tests live in tests/ and use typer.testing.CliRunner with fake ApiClient and SessionStore implementations defined in tests/conftest.py. No real network calls are made.

---

License

MIT