ApplyBot

AI-assisted job discovery and application tracking through Telegram, with configurable filtering, resume matching, Supabase persistence, and Google Sheets reporting.

ApplyBot does not submit applications or store LinkedIn credentials. It helps discover, evaluate, and track opportunities while keeping the final application step under the user's control.

Highlights

• Aggregates roles from public job APIs and configured Greenhouse, Lever, and Ashby boards
• Applies deterministic role, experience, location, and score gates before using LLM quota
• Scores jobs against a candidate profile and produces resume-tailoring suggestions
• Routes AI requests across Gemini, NVIDIA NIM, and Groq with quota-aware fallbacks
• Deduplicates jobs and resumes safely after interrupted runs
• Tracks application stages through Telegram and logs qualified roles to Google Sheets
• Redacts sensitive values from operational logs
• Includes 200+ offline tests

Architecture

Job APIs and ATS boards
          |
          v
 Discovery -> normalization -> deduplication -> eligibility filters
                                              |
                                              v
                                    optional semantic matching
                                              |
                    +-------------------------+-------------------------+
                    |                         |                         |
                 Supabase                Google Sheets              Telegram
              source of truth             reporting              interaction

Core modules:

• app/jobs/ — discovery, normalization, filtering, scoring, and deduplication
• app/llm/ — provider routing, rate-limit handling, and resume analysis
• app/telegram/ — commands and application workflow
• app/sheets/ — formatted workbook output
• supabase/migrations/ — database schema
• tests/ — offline unit and integration coverage

Quick start

Prerequisites

• Python 3.10+
• A Supabase project
• A Telegram bot
• Optional Google Sheets and LLM-provider credentials

Installation

git clone https://github.com/ad1820/ApplyBot.git
cd ApplyBot
python -m venv .venv

Activate the environment and install dependencies:

pip install -r requirements.txt

Copy .env.example to .env, then add the credentials required for the integrations you want to enable. See HOW_TO_RUN.md for the complete configuration walkthrough.

Run discovery:

python scripts/run_discovery.py

Start the Telegram interface:

python scripts/telegram_polling.py

Telegram commands

Command	Purpose
/today, /jobs, /job <id>	Browse discovered jobs
/done <id>, /skip <id>	Record an application decision
/status <id>, /applied, /stats	Review application progress
/interview <id>, /offer <id>	Move an application through the pipeline
/setresume <text>	Store the master resume
/resume <id>	Analyze resume fit for a job
/approveresume <id>	Approve a tailored-resume draft

Testing

The test suite is designed to run without live external services:

pytest

Responsible use

ApplyBot is an assistant, not an autonomous applicant. Review generated resume content for accuracy, verify job details at the original source, and submit applications manually. Keep credentials in local environment variables and never commit a populated .env file.

Documentation

• Daily workflow
• Complete setup guide
• Architecture and engineering notes