Skip to content

Agnostos

Resume-driven outreach for research internships. Agnostos is a free, open-source command-line tool that helps students find research internships:

  • Who should I write to? It ranks a professor list against your resume locally, for free, or builds the list for you from a research topic using the OpenAlex paper index.
  • What have they worked on lately? It pulls each professor's recent first- and last-author papers.
  • What do I say? One language-model call per professor returns an acceptance estimate and a short, specific, honest draft. Common AI tells are stripped out and unverifiable claims are flagged.
  • Is the address right? It checks every address and can find the one each professor publishes on their own page.
  • When do I send it? Slowly, from your own inbox, during each professor's working hours.
  • Did they answer? It tracks replies, auto-replies and bounces, and drafts follow-ups for you to approve.

Get it

Source on GitHub Technical report (PDF)

git clone https://github.com/Avinashistoocool/Agnostos-v2
cd Agnostos-v2
pip install -r requirements.txt
cp .env.example .env                       # Windows: copy .env.example .env
python -m outreach.cli profile             # check resume parsing
python -m outreach.cli run --mode list     # score + draft into your xlsx
python -m outreach.cli send --dry-run      # preview the emails

The full command reference is in the README.

Documentation
Setup: Windows · macOS · Linux Step-by-step installation and daily use for each operating system
FAQ Common questions about costs, sending, replies and safety
Diagnostics Health checks and fixes for every common error message
Google Calendar One-time setup for meeting-slot checks and calendar holds
Technical report (PDF) Architecture and every scoring rule: matcher, discovery ranking, cost estimate, AI-tell score, address confidence, scheduling

Note

Agnostos drafts; you decide. Nothing is sent without --live, and every draft is written to your spreadsheet first so you can read and edit it. The acceptance percentage is a model's estimate for ordering your list, not a statistical probability.