Stupid Emails
https://github.com/farzaa/stupid-emails- Category
- Productivity
- Rank
- No. 2190Tools index
Previous survey · No. 2165 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- farzaa
- GitHub
- 90 stars
- Date
About
A simple Python script to tag your Gmail inbox with AI.
What it does
It reads Gmail messages, cleans and shortens their contents, then asks an OpenAI model to choose a personalized category. It applies a fixed set of labels shaped by a prompt containing the user’s bio, priorities, and examples.
Why it's ranked here
This is a compact, inspectable automation for people willing to tune their own inbox rules. Its strongest idea is prompt-driven personal priority, not generic spam detection. Setup friction, broad Gmail permissions, and brittle model output handling keep it firmly in hobby-project territory.
What's good
The classifier considers sender, subject, cleaned content, and message date. It handles plain text and HTML, strips common newsletter clutter, and caps content at 500 characters to reduce model input. Credentials can come from local files or environment variables, which supports local and hosted worker setups.
Tradeoffs
Every processed email is sent to OpenAI, despite the README’s narrower claim that nothing is saved. Gmail access includes both read and modify scopes. Classification depends on valid JSON from one dated model identifier, with no retry or schema enforcement. Aggressive cleaning and truncation can remove context needed for a correct label.
How to use it well
Use it as a personal inbox triage experiment when you can maintain a detailed prompt and verify its labels. Start with narrow categories and review mistakes before trusting automation. It does not provide a polished application, general email client, shared team workflow, or evidence of tested production reliability.
Technical notes+
The repository exposes a script-style worker in tag.py, launched locally with python tag.py or through Procfile. runtime.txt specifies Python 3.11, while requirements.txt lists Google OAuth, Gmail API, OpenAI, tokenization, HTML parsing, and scheduling packages. tag.py loads OPENAI_API_KEY or secrets.json, accepts Gmail credentials through environment variables or credentials.json, and caches local OAuth state in token.pickle. It requests Gmail readonly and modify scopes, parses MIME bodies recursively, cleans HTML with BeautifulSoup, truncates content to 500 characters, and calls gpt-5-2025-08-07 through chat completions. prompt.txt defines six fixed labels and extensive personal classification guidance. The supplied tag.py text is truncated, so later message-querying and label-application mechanics cannot be verified here.
Observed
- Primary language
- Python
- Runtime
- Python 3.11 is specified
- Interface
- Command-line script and Procfile worker
- Install surface
- Git clone followed by pip installation from requirements.txt
- Platforms and services
- Gmail API and OpenAI API
- Configuration
- Prompt text, API secrets, and Gmail OAuth credentials are supplied separately
- License
- No license file or license terms appear in the provided repository text
Read from README.md, requirements.txt, tag.py, Procfile, prompt.txt, runtime.txt, secrets.template.json, secrets_template.json.
What it can do
Connect to Gmail inbox
Gmail account credentials → Access to Gmail inbox
Analyze email content with AI
Email messages from Gmail inbox → AI-generated analysis or categorization
Generate tags for emails
Email content and AI analysis → Relevant tags or labels
Apply tags to Gmail messages
Generated tags and target emails → Tagged emails in Gmail inbox
Process multiple emails in batch
Gmail inbox with multiple untagged emails → Bulk-tagged email collection
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.