BUILD WITH JEVROUTER

From example to application.

Connect a model, understand the result, and keep your costs under control.

Chat integration

Use an OpenAI-compatible model API with Python

Connect the OpenAI Python SDK to Jevrouter, make a fixed-model chat request, inspect usage and handle request failures without duplicate generations.

Read the guide →

Streaming

Stream a model response with JavaScript

Read a chat completion stream using server-side JavaScript, display text chunks, capture usage and avoid retrying a partially delivered answer.

Read the guide →

Images

Generate and save an image with Python

Call Jevrouter's native image API from Python, read SSE completion events, save a PNG and understand image usage receipts and balance reservations.

Read the guide →

Production basics

Control API budgets and retry safely

Understand Jevrouter project balances, key budgets, per-request cost ceilings, idempotency conflicts and how to recover from unknown API outcomes.

Read the guide →

Model selection

LLM routing: choose models and measure the tradeoffs

Learn how Jevrouter uses Jev task signals, model allowlists and budget constraints, and how to evaluate routing against fixed-model and rules-only baselines.

Read the guide →

Put numbers behind your model choice.

Estimate token and image costs with the current catalog, then test the models against your own acceptance criteria.

Cost calculator →