From request to workflow
Jev API guides and agent workflow examples
Learn Jev through OmniaKey: API foundations, model routing, and complete workflows with GPT-6.1 Sol, Claude Sonnet 5.5 and Claude Opus 5.5.
阅读中文版New to Jev? Start with your first request, then learn the three decision types. Already calling the API? Pick the workflow that matches your task. These guides are independently published by Jev AI; OmniaKey provides accounts and API access.
01 / API foundations
Jev API quickstart with OmniaKey
Make your first Jev API request with an OmniaKey key. Follow curl and Node.js examples, inspect the response, and troubleshoot authentication errors.
Read guideAPI conceptsJev Choice, Score and Noul explained
Choose the right Jev question type. Understand option probabilities, fractional scores, Noul values and confidence with a worked refund-review example.
Read guide02 / Agent workflows
Each guide includes a complete request, answer interpretation, evaluation cases and a matching playground preset to help you assess your own task.
LLM model routing with Jev: a practical example
Build a bounded LLM router with Jev. Describe approved models, compare task constraints, handle escalation and evaluate cost against a fixed-model baseline.
Read guideTool reviewAI agent tool-call guardrails with Jev
Review proposed agent tool calls with Jev using action choices and a risk scale. Keep authorization, approval thresholds and execution checks in your code.
Read guideEvidence reviewResearch verification with Jev: check claims against evidence
Check whether supplied evidence supports a research claim with Jev. Handle insufficient, contradictory and stale evidence without treating the model as a web search engine.
Read guideCompletion reviewAI agent task completion checks with Jev
Review an agent's work against its goal, verification and known gaps. Use Jev to distinguish done, verify more, continue work and blocked outcomes.
Read guide03 / Jev × Generation models
Check answers after Sol drafts them, route tickets before Sonnet writes, and review operations proposed by Opus. Each guide includes a complete two-model script and a native Jev snapshot for single-step testing.
GPT-6.1 Sol + Jev: generate an answer, then check its evidence
Connect GPT-6.1 Sol and Jev through OmniaKey. Generate a support draft, check every sentence against original evidence, and handle one revision in Node.js.
Read guideJev × Sonnet · Ticket routingClaude Sonnet 5.5 + Jev: triage support tickets before drafting replies
Use Jev to route a support ticket, then Claude Sonnet 5.5 to draft a reply or request details. Follow a complete OmniaKey example with identity checks.
Read guideJev × Opus · Operation reviewClaude Opus 5.5 + Jev: review an operation proposal before execution
Let Claude Opus 5.5 propose an operation and Jev review intent and safeguards. Build an OmniaKey workflow with tool validation and explicit approval branches.
Read guideFor current account and API policies: OmniaKey documentation.