Skip to content

README.md

Latest commit

 

History

History
92 lines (66 loc) · 2.81 KB

README.md

File metadata and controls

92 lines (66 loc) · 2.81 KB

cold-send0r-bot

CLI tool that automates personalized cold outreach emails. Scrapes company websites, generates tailored emails via LLM, and sends them over SMTP.

contacts.json ──> scrape ──> generate (LLM) ──> review ──> send

Quick Start

git clone https://github.com/dante4rt/cold-send0r-bot.git
cd cold-send0r-bot
go build -o send0r .

# Create your config files from templates
./send0r init

Edit the generated files:

File What to change
config.yaml Your name, email, links, SMTP/LLM settings
resume.txt Your background summary (fed to the LLM)
contacts.json Target contacts with company URLs
.env API keys: OPENROUTER_API_KEY, SMTP_USERNAME, SMTP_PASSWORD

Then run:

./send0r pipeline --dry-run    # generate emails without sending

Tip

Review output/emails.json before removing --dry-run.

Commands

Command Description
init Copy example configs to working files
pipeline Full flow: scrape + generate + send
scrape Scrape company websites only
generate Generate emails from scraped data
send Send previously generated emails

All commands support --verbose and --config <path>.

Config

See config.example.yaml for all options. Key sections:

Section Controls
sender Your name, email, and links (GitHub, etc.)
scraper Provider (colly/firecrawl), rate limits
llm Model, temperature, token limit via OpenRouter
smtp Host, port, credentials (via env vars)

Contact format

Each entry in contacts.json:

{
  "email": "hiring@example.com",
  "name": "Jane Doe",
  "company": "Example Corp",
  "role": "Engineering Manager",
  "url": "https://example.com"
}

How It Works

  1. Scrape -- Fetches each contact's company URL (Colly + optional Rod headless fallback)
  2. Generate -- Sends company content + your resume to an LLM via OpenRouter, producing a personalized subject + body
  3. Send -- Delivers emails over SMTP with optional PDF attachments and rate limiting

Important

The LLM uses your sender.links from config. No personal data is hardcoded in the source.

Requirements

  • Go 1.21+
  • OpenRouter API key
  • SMTP credentials (Gmail app password, etc.)

License

MIT