یک API برای همهٔ مدل‌ها

Connect this workspace to IranRouter (OpenAI-compatible LLM gateway).

1. In Cursor Settings → Models, enable "Override OpenAI Base URL".
2. Set Base URL to: https://iranrouter.net/api/v1
3. Set API Key to my IranRouter key (env: IRANROUTER_API_KEY).
4. Add to .env.example and .env:
   IRANROUTER_API_KEY=ir-your-key-here
   OPENAI_BASE_URL=https://iranrouter.net/api/v1
5. Keep your existing model names (e.g. deepseek-chat, gpt-4o-mini) — only base URL and API key change.
6. Verify with a quick test request to https://iranrouter.net/api/v1/chat/completions.

Docs: https://iranrouter.net/docs

Wire this project to use IranRouter as the LLM backend (OpenAI-compatible API).

Requirements:
- Base URL: https://iranrouter.net/api/v1
- Auth: Authorization: Bearer $IRANROUTER_API_KEY
- Models: keep provider-native names (deepseek-chat, gpt-4o-mini); prefixed slugs still work

Tasks:
1. Add IRANROUTER_API_KEY to .env and .env.example.
2. Configure the OpenAI client (or fetch) to use base URL https://iranrouter.net/api/v1 instead of api.openai.com.
3. Do not rename models — only swap base URL and API key.
4. Add a small smoke test or script that sends one chat message and prints the reply.
5. Document the setup in README if one exists.

IranRouter docs: https://iranrouter.net/docs

Configure this project to route all LLM calls through IranRouter (OpenAI-compatible gateway).

Settings:
- API base: https://iranrouter.net/api/v1
- API key env var: IRANROUTER_API_KEY
- Example models: deepseek-chat, gpt-4o-mini (keep provider-native names)

Please:
1. Add IRANROUTER_API_KEY to environment configuration.
2. Point the OpenAI-compatible client base URL to https://iranrouter.net/api/v1.
3. Keep existing model names; only swap base URL and API key.
4. Run a quick connectivity check (GET https://iranrouter.net/api/v1/models or one chat completion).
5. Keep existing app structure; only change the LLM provider wiring.

Reference: https://iranrouter.net/docs

Use IranRouter instead of OpenAI directly. IranRouter is a drop-in OpenAI-compatible API.

Configuration:
  base_url = "https://iranrouter.net/api/v1"
  api_key  = process.env.IRANROUTER_API_KEY   # Bearer ir-…

Example (Node):
  import OpenAI from "openai";
  const client = new OpenAI({
    apiKey: process.env.IRANROUTER_API_KEY,
    baseURL: "https://iranrouter.net/api/v1",
  });
  const res = await client.chat.completions.create({
    model: "deepseek-chat",
    messages: [{ role: "user", content: "Hello" }],
  });

Example (Python):
  from openai import OpenAI
  client = OpenAI(
      api_key=os.environ["IRANROUTER_API_KEY"],
      base_url="https://iranrouter.net/api/v1",
  )

List models: GET https://iranrouter.net/api/v1/models
Docs: https://iranrouter.net/docs