Research-aware routing
Compare claims, sources, or alternatives and return the next approved route instead of another paragraph.
Decision model · Perplexity
Turn research context into a decision.
PPLX Decider is a strong fit when an agent must compare evidence, rank alternatives, or route a research task without making your application parse prose.
Use the exact model ID in the request. Credits follow the same Jev rule as the other decision models.
A decision boundary
perplexity/pplx-decider-v1-27bOne state. Three useful outputs.
State
Two vendors meet the reliability requirement. Vendor B is cheaper, but has weaker recent evidence.
Choice
Route to vendor A
Score
0.78 confidence
Noul
Escalate if evidence is stale
The output is typed for your router, evaluator, or review queue.
Input
State + typed questions
Output
Choice · score · noul
Billing
Same as Jev
Endpoint
POST /v1/systemone
Why this model
A clean boundary for evidence-led agents.
Keep retrieval and actions in your application. Ask PPLX Decider for the bounded judgment in between.
Typed answers
Choice, score, and noul map directly to application branches, confidence gates, and review policies.
Small decision surface
Send only the state and criteria needed for this decision so the surrounding workflow stays inspectable.
Where it fits
Use it when the question is about evidence.
PPLX Decider is most useful after your application has gathered context and before it commits to a route.
PatternUse it forGood boundary
Evidence ranking
Compare sources, candidates, or claims
Return the best-supported option
Research triage
Decide whether a result is ready to use
Gate downstream drafting or action
Uncertainty handling
Detect weak or conflicting evidence
Send the case to human review
Call the model
One request, one typed decision.
The Jev API keeps the wire format stable across supported decision models. Change only the model ID when you want to compare behavior.
- Set model to the exact model ID shown below.
- Send the current state as a compact string or JSON-serializable value.
- Ask one or more typed questions and handle the returned answer in application code.
The model ID is perplexity/pplx-decider-v1-27b.
Use choice for a closed set of options, score for a bounded value, and noul for a yes/no-style boundary.
Validate confidence and review conditions before taking an irreversible action.
Example request
curl -X POST https://thejevai.com/v1/systemone \
-H "Authorization: Bearer <API_KEY>" \
-H "Content-Type: application/json" \
-d '{
"model": "perplexity/pplx-decider-v1-27b",
"state": "A customer request needs a safe next step.",
"questions": {
"route": {
"type": "choice",
"instructions": "Choose the correct next step.",
"criteria": {
"continue": "Continue the approved workflow",
"review": "Send the case for human review"
}
}
}
}'Keep the API key on your server. The Playground uses the same request shape through your signed-in session.
Build with it
Make research workflows easier to operate.
Start with a decision that has a clear owner, a small set of options, and an observable fallback.
Source selection
Choose which source or candidate should move into the next step.
Answer readiness
Gate an answer when evidence is incomplete, stale, or contradictory.
Analyst escalation
Turn uncertain research cases into explicit review work instead of silent guesses.
FAQ
Before you ship the first decision.
How do I select PPLX Decider?
Set model to perplexity/pplx-decider-v1-27b in the POST /v1/systemone request. The Playground exposes the same model in its model selector.
Does it use the same billing rule as Jev?
Yes. The supported decision models use the same Jev-based credit rule. Check your own production-like cases because latency, availability, and provider behavior can change.
Can it execute the final action?
No. Treat the response as a typed recommendation. Your application should own tool calls, writes, permissions, and human review.
References
Keep the model and API boundary visible.
OpenRouter model page
Check current availability and provider-specific model details before production rollout.
Jev API documentation
Use the shared request, response, authentication, and error contract.
Your evaluation set
Measure routing accuracy, confidence calibration, and review rate with representative cases.
Start with a bounded question
Give your research agent a decision boundary.
Try PPLX Decider in the Playground, then move the same request to your server when the output is stable.