Pick the intelligence tier and maximum price yourself. CPEN routes inside that ceiling.
Cheap, Controllable, Cost Efficient, ConfigurableControl intelligence and top price for every AI request.
CPEN gives developers one OpenAI-compatible API, real-time price checks, hard USD caps, provider-group status, and automatic route selection for extraction, cleanup, batch jobs, and mini-chat.
Live caps, provider-group health, and route aliases in one request path.
A route is selected after intelligence, status, and your price ceiling are checked.
CPEN keeps the developer-facing surface stable while the provider group can change behind the alias to stay inside the quality level and maximum price you set.
Control the ceiling before the model choice
The product is optimized for teams that want to define acceptable intelligence, maximum price, fallback behavior, and batch throughput before dispatch.
Router catalog
Use stable CPEN aliases first. Pin provider groups later when you know exactly what intelligence level, price ceiling, uptime, region, or data policy you need.
| Route | Best live model | Provider group | Input | Output | Total + fee | Candidates | Status |
|---|---|---|---|---|---|---|---|
| cpen/auto-chatChat | gpt-5.4-mini | Naapi / GPT | $0.0206 | $0.1238 | $0.1516 | 107 | available |
| cpen/auto-extractExtract | gemini-2.5-flash | Yunwu / Gemini | $0.0274 | $0.1094 | $0.1436 | 31 | available |
| cpen/auto-cheapBatch | deepseek-v3-mini | DuckCoding / Value | $0.0089 | $0.0351 | $0.0462 | 62 | available |
| cpen/pin-providerPolicy | user selected | Pinned group | cap based | cap based | strict | 0 | planned |
Drop-in quickstart
Swap the base URL, keep the OpenAI-style request shape, and set per-request intelligence and USD routing caps.
const client = new OpenAI({
baseURL: "https://cpenrouter.space/v1",
apiKey: process.env.CPEN_API_KEY
});
await client.chat.completions.create({
model: "cpen/auto-extract",
messages: [{ role: "user", content: row }],
routing: {
currency: "USD",
max_total_per_1m: "0.15"
}
});