Classify email replies for workflow automation. Approval, rejection, question, escalation, noise. Deterministic rules. No ML. No data leaves your machine.
Email handling for automated workflows is broken.
Following up with a vendor, getting approval from a client, handling support escalations. Most teams end up with one of three bad solutions: polling a shared Gmail inbox manually, using a transactional email API that is blind to replies, or building a custom email server and spending weeks on deliverability.
Existing tools are expensive and complex.
You need something that just classifies the reply and gets out of the way. No $50/month subscription, no complex setup.
Feed it a plain-text email reply. Get back structured classification:
Sender approves or agrees. Action: proceed.
Sender rejects or declines. Action: stop.
Sender asks a question. Action: respond.
Sender escalates to higher authority. Action: escalate.
Not actionable. Action: ignore.
$ echo '{"text":"I approve this. Let'"'"'s proceed."}' | python handler.py
{"bucket": "approval", "confidence": 0.9, "action": "proceed"}
Deterministic keyword matching with priority ordering. No model, no API call, no data leaves your machine.
An LLM classifying email replies is a black box you cannot audit.
When a tool says "this is an approval", you need to know why. A rule you can read is a rule you can trust, debug, and improve.
EUR29 one-time. Runs locally. No subscription, no cloud, no data collection.
Buy now - EUR29