Migrate from DeepSeek Direct to Railwail

TL;DR — Switch in Under 5 Minutes

• DeepSeek's API is OpenAI-compatible — change baseURL only
• DeepSeek V3, R1, Coder, Math, R1-Distill all available on Railwail
• EU-hosted endpoint — addresses common compliance concerns about routing requests through China
• EUR billing with VAT receipts
• Plus access to Claude, GPT-4o, Gemini, Llama through the same key

Why Move Off DeepSeek Direct?

DeepSeek's frontier-quality models (V3, R1) at low prices are compelling. The trade-off many EU teams encounter: the deepseek.com direct API routes requests through China, which can be a regulatory or compliance blocker. Railwail hosts DeepSeek's open-weight models on EU infrastructure (Hetzner Frankfurt) so you get the same model with EU data residency.

Step 1 — Get a Railwail API Key

Sign up at railwail.com and generate a key.

Step 2 — Change Base URL

TypeScript / JavaScript

Before (DeepSeek direct):

import OpenAI from "openai";

const ds = new OpenAI({
  apiKey: process.env.DEEPSEEK_API_KEY,
  baseURL: "https://api.deepseek.com/v1",
});

const res = await ds.chat.completions.create({
  model: "deepseek-chat",
  messages: [{ role: "user", content: "Hello" }],
});

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.RAILWAIL_API_KEY,
  baseURL: "https://api.railwail.com/v1",
});

const res = await client.chat.completions.create({
  model: "deepseek-v3",
  messages: [{ role: "user", content: "Hello" }],
});

Python — Reasoning with R1

from openai import OpenAI

client = OpenAI(
    api_key=os.environ["RAILWAIL_API_KEY"],
    base_url="https://api.railwail.com/v1",
)

resp = client.chat.completions.create(
    model="deepseek-r1",
    messages=[{"role": "user", "content": "Prove that the square root of 2 is irrational."}],
)
print(resp.choices[0].message.reasoning_content)  # chain-of-thought
print(resp.choices[0].message.content)             # final answer

cURL

curl https://api.railwail.com/v1/chat/completions \
  -H "Authorization: Bearer $RAILWAIL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v3",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

API Endpoint Mapping

Model Mapping

Pricing Comparison (per 1M tokens, May 2026)

Why Railwail Over DeepSeek Direct

• EU hosting — data residency in Frankfurt, no routing to Chinese servers
• EUR billing with VAT receipts
• Same OpenAI-compatible API
• Adds Claude, GPT-4o, Gemini for the rest of your stack
• Per-key rate limits and spend caps
• Built-in playground for R1 reasoning chain inspection

FAQ

Is my data really not sent to China?

Correct. Railwail hosts DeepSeek's open-weight models (V3, R1, R1-Distill, Coder) on Hetzner Frankfurt GPU instances. Your requests and the model weights run entirely within Hetzner's EU data centres. The original DeepSeek company's servers in China are not involved.

Do I get the same model quality?

Yes — same open-source weights from huggingface.co/deepseek-ai. Output distribution is identical at matched sampling params.

Is reasoning_content returned for R1?

Yes. The chain-of-thought reasoning is available in choices[0].message.reasoning_content. The final answer is in choices[0].message.content.

Does the prompt caching discount work?

Yes. DeepSeek's automatic context caching is preserved — repeated prompt prefixes trigger the cache-hit pricing tier.

What about the FIM (fill-in-the-middle) mode for deepseek-coder?

POST /v1/completions with deepseek-coder and prefix/suffix params triggers FIM.

Next Steps

• Sign up at railwail.com
• Generate an API key
• Update baseURL to https://api.railwail.com/v1
• Switch model IDs (deepseek-chat → deepseek-v3, deepseek-reasoner → deepseek-r1)
• Read the reference at railwail.com/docs
• Compare pricing at railwail.com/pricing