Unofficial field guide to TypeSafe AI's Jev
Practical Jev use cases, with working code
Jev is a "System One" model: you send a state (an email, a log line, a ticket, any JSON) plus a list of typed questions. It answers every question in parallel with a Choice, a Score, or a Noul (yes/no) value, each backed by probabilities, instead of generated text. Your code branches on those answers and uses confidence to decide when to act and when to escalate.
{
"subject": "Charged twice this month",
"body": "Hi, I see two charges of $49 on my
card for September. Can you refund one?"
}queuechoicebillingfrustrationscore0.45is_outage_reportnoul0.02Jev use cases
Content Moderation and Injection Screening with Jev
Screen user messages and retrieved passages before they reach your LLM. One Jev request returns a policy Choice plus Noul checks for jailbreaks and injections.
Email Triage and Priority Scoring with Jev
Sort a shared inbox with one Jev request. A Choice sets the category, three Score questions rate urgency, sender signals and effort, and code weights them.
Hierarchical Product Categorization with Jev
Place marketplace listings in a taxonomy with thousands of leaves using cascading Jev Choice questions, a small beam, and confidence-based parent back-off.
Intent Routing for Chatbots and Agents with Jev
Route chatbot and agent messages across 60+ intents with one Jev Choice built from your intent registry, a none option, and a Noul gate. Code and cost math.
Invoice and Document Classification with Jev
Classify OCR text from an accounts payable inbox with one Jev Choice question, flag reminders and past-due notices with Nouls, keep amounts and dates in code.
Lead Scoring with Jev Score Questions
Score inbound demo requests with four atomic Jev Score questions for fit, intent, seniority and urgency, then combine them with weights you control in code.
LLM Output QA and Agent-Run Review with Jev
Check generated answers and agent transcripts before they ship with Jev Noul and Score questions, then gate release on thresholds. Code and cost math.
Log Anomaly Classification with Jev
Classify flagged log clusters into cause categories with one Jev Choice and speculative Noul checks, while counting and rate math stay in your own code.
News Relevance Filtering with Jev
Filter headlines against a client profile with Jev Noul and Choice questions, one request per article, run concurrently. Includes code and cost estimates.
Search Result Re-Ranking with Jev
Re-rank BM25 or vector search results with one Jev Score question per query and candidate pair, run concurrently, then sort in code. Python, TypeScript, costs.
Security Alert Triage with Jev
Triage SIEM and EDR alerts with Jev Score and Choice questions, rank the analyst queue, and keep host isolation behind a high confidence bar and a human.
Support Ticket Routing with Jev
Route support tickets to the right team with one Jev Choice question, then gate auto-assignment on confidence. Python and TypeScript code plus cost math.
Guides
What is Jev?
System One vs LLMs, the three primitives, confidence, pricing, and limitations.
Jev vs LLMs for classification
Cost and latency math, and an honest list of when an LLM is the better tool.
Confidence thresholds
How to gate automated actions on confidence and tune the thresholds on your data.
Go to the source
This is an unofficial community site. API details change, so treat theofficial TypeSafe documentation as the source of truth, starting with theHTTP API reference and the list of known Jev limitations.