Documentation

Build around decisions

The Jev-compatible API supports routing, classification and scoring from text. Choose Peer-to-Peer for elastic capacity or EU for dedicated EU processing. The API accepts context and up to three typed text questions for s1-pro, or one for s1-fast and s1-vision, per request, then returns a structured answer. Public routes and schemas follow the current API contract and are published as OpenAPI 3.1.

01 / FIRST REQUEST

Quickstart

See request/response samples and illustrative curl, Python, JavaScript, TypeScript, Rust, C# and Java examples.

Open the quickstart →

02 / CONTRACT

API reference

Browse the static OpenAPI 3.1 contract snapshot and the request, answer and error schemas.

Open API reference →

03 / DIFFERENCES

Migrate from Jev

Move the base URL and credentials, then review the documented differences and retry behavior.

Read migration notes →

04 / PATTERNS

Examples

Typed routing, moderation guardrails and bounded image classification patterns.

Browse examples →

ROUTING / s1-llm-auto-router

LLM auto-router model

Classify a request for an LLM router in one call: category, difficulty, stakes and four flags.

Read the router model docs →

05 / PROCESSING

Regions & data

Choose Peer-to-Peer or optional EU processing; understand current routing, payload retention, usage-history settings and idempotency.

Read data handling →

06 / OPERATIONS

Errors

Use request IDs, retryable errors, and `Retry-After` without repeating a charge.

Review error codes →
Contract note

Examples are illustrative. The OpenAPI document is the source of truth for routes and schemas. Check the OpenAPI document again before relying on any route.

https://api.system1models.ai keeps working.