Bilingual support
deepseek-chat
Cost-efficient drafts; spot-check with native speakers.
Europe · bilingual + controls
Budgeted keys per squadEuropean squads usually need bilingual quality checks, clear key ownership and prepaid spend caps before a wider rollout. Keep one Base URL, separate keys per team, and only promote Live DeepSeek routes.
Europe pilot plan
| Step | Action | Success signal |
|---|---|---|
| Define languages | Pick 2–3 live product languages and write pass/fail rubrics before calling any model. | Rubric exists for tone, factuality and refusal cases. |
| Split keys | Issue one budgeted key per squad (support, eng, research). | Wallet logs show attributable usage by team. |
| Promote carefully | Keep production on Live IDs only; document Integrating entries as roadmap. | No Integrating ID appears in prod config. |
Workload matrix
deepseek-chat
Cost-efficient drafts; spot-check with native speakers.
deepseek-v4-pro
Stronger multi-step answers on the same evidence pack.
deepseek-chat → V4 Pro escalate
Use Pro only when chat fails the test suite.
Live prices
| Model | Billable price |
|---|---|
| DeepSeek Flash deepseek-chat | $0.3 input / $1.2 output per 1M |
| DeepSeek V4 Pro deepseek-v4-pro | $1.32 input / $3.96 output per 1M |
Integrating catalog families are market context only. Confirm Live status on /models before production.
Run identical tickets in English and a local language. Score accuracy and tone before choosing a default Live route.
Named keys with daily caps make it easier to explain AI spend in monthly reviews without opening every vendor console.
Keep customer prompts in approved environments, log request IDs (not secrets), and treat catalog status as part of change control.
Migration
import OpenAI from "openai";
const client = new OpenAI({
apiKey: process.env.SMTOKEN_API_KEY,
baseURL: "https://thesmartoken.com/v1",
});
const completion = await client.chat.completions.create({
model: "deepseek-chat",
messages: [
{
role: "user",
content: "Draft a English reply for our Europe support pilot.",
},
],
});No. The Base URL remains https://thesmartoken.com/v1. Regional pages describe workloads and operating habits, not separate model clusters.
Yes. Keep the SDK, set baseURL, authenticate with a SmarToken key, and select a Live model ID.
deepseek-chat and deepseek-v4-pro. Re-check /models and /pricing before procurement quotes.
Treat them as catalog signals during wiring, not production call targets.