Can I reuse my existing DeepSeek code with GotoRamp?

Yes, if it sends OpenAI-compatible requests, for example through the OpenAI SDK. Change the base URL to https://api.gotoramp.ai/v1 , use your GotoRamp key, and keep a model ID that GET /v1/models returns.

Which DeepSeek model ID should I start with?

Start with deepseek-chat for general chat. GET /v1/models lists the other DeepSeek models open to your account, including reasoning models where available.

Does DeepSeek support tool calling through GotoRamp?

It depends on the model version. Where the console shows tool calling for a version, you pass tools and tool_choice in the usual OpenAI-compatible format.

How is DeepSeek usage billed?

Per token, with input and output tokens billed at separate rates, from a prepaid balance in US dollars. Rates are shared with early-access accounts, and failed requests are not billed.

Do I need a DeepSeek account as well?

No. You call DeepSeek models with a GotoRamp key, pay from your GotoRamp balance and receive invoices from GotoRamp LLC in USD.

Models · Text

DeepSeek API for businesses in Southeast Asia

DeepSeek's chat and reasoning models, made by DeepSeek in China, are opening with early access on GotoRamp, a US-run, OpenAI-compatible API for businesses in Singapore, Malaysia, the Philippines and Thailand.

At a glance

DeepSeek on GotoRamp.

If your code already calls an OpenAI-compatible endpoint, moving it to DeepSeek on GotoRamp takes a new base URL, a new key and a model name.

GotoRamp LLC is a US company in Mountain View, California, and isn't affiliated with DeepSeek. We offer a model only once we hold written terms for it through a channel that allows our use in each market.

MakerDeepSeek (China)
TypeText: chat and reasoning models
How you call itOpenAI-compatible POST /v1/chat/completions through GotoRamp
Example model IDdeepseek-chat
BillingPer input and output token, at separate rates, from a prepaid USD balance
StatusOpening with early access
CountriesSingapore, Malaysia, the Philippines and Thailand first

Use cases

What teams use it for.

DeepSeek models are widely used for coding, analysis and agent workflows. Test on your own data before you ship: results vary by task, language and model version.

Coding help inside your tools

Review comments, test scaffolding, refactoring suggestions and plain-English explanations of unfamiliar code, wired into your editor, repository bot or internal portal. A developer still approves what gets merged.

Analysis a person can check

Turn support tickets, sales notes or survey answers into structured findings. For questions that take several steps, try a reasoning model and ask it to lay out its steps so a reviewer can audit the result.

Agent workflows

Multi-step tasks that touch your own systems: look up an order, draft the reply, open a ticket. Your code runs each tool; the model picks the next one, on versions that support tool calling.

Call it with the OpenAI SDK

Point the SDK at GotoRamp.

Install the OpenAI SDK, set base_url to https://api.gotoramp.ai/v1, read your key from GOTORAMP_API_KEY and send deepseek-chat as the model.

The endpoint opens with your early-access invitation. Every field, error code and streaming detail is in the chat completions reference, and the OpenAI SDK guide covers moving an existing project.

Python

# pip install openai
import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.gotoramp.ai/v1",
    api_key=os.environ["GOTORAMP_API_KEY"],
)

review = client.chat.completions.create(
    model="deepseek-chat",
    messages=[
        {"role": "system", "content": "You review Python code. Point out bugs first, then style. Be brief."},
        {"role": "user", "content": "def average(xs):\n    return sum(xs) / len(xs)"},
    ],
    temperature=0.2,
)

print(review.choices[0].message.content)
print(review.usage.prompt_tokens, review.usage.completion_tokens)  # billed per token

Things to know

Before you build on DeepSeek.

Five facts that decide whether a DeepSeek model fits your product.

Features vary by versionTool calling, JSON output and context length differ between DeepSeek versions. Check GET /v1/models and the console before you depend on one.
Chat or reasoningChoose per request. Run your own prompts through both and compare the usage block, because output tokens are billed as well as input tokens.
Content rulesDeepSeek's own content policies apply alongside our Acceptable Use Policy, so some prompts may be refused. A refused request returns 403 content_policy and isn't billed.
Where data is processedThe processing location for each model is listed in our Privacy Notice before accounts open.
TrainingWe don't use your prompts, files or outputs to train models.

FAQ

DeepSeek questions.

Something missing? Ask us; we reply within one business day.

Can I reuse my existing DeepSeek code with GotoRamp?: Yes, if it sends OpenAI-compatible requests, for example through the OpenAI SDK. Change the base URL to https://api.gotoramp.ai/v1, use your GotoRamp key, and keep a model ID that GET /v1/models returns.
Which DeepSeek model ID should I start with?: Start with deepseek-chat for general chat. GET /v1/models lists the other DeepSeek models open to your account, including reasoning models where available.
Does DeepSeek support tool calling through GotoRamp?: It depends on the model version. Where the console shows tool calling for a version, you pass tools and tool_choice in the usual OpenAI-compatible format.
How is DeepSeek usage billed?: Per token, with input and output tokens billed at separate rates, from a prepaid balance in US dollars. Rates are shared with early-access accounts, and failed requests are not billed.
Do I need a DeepSeek account as well?: No. You call DeepSeek models with a GotoRamp key, pay from your GotoRamp balance and receive invoices from GotoRamp LLC in USD.

Other models

One account, one balance.

Switching families is a change to the model field. Compare every model.

QwenAlibaba Cloud · chat, reasoning and vision-language models
KimiMoonshot AI · long-context and agent-oriented models
GLMZhipu AI (Z.ai) · general-purpose models used for coding and agents
Doubao SeedByteDance Seed · general chat and multimodal models
SeedanceByteDance · video generation, billed per second

Start with a small balance. Scale when it works.

We're onboarding early-access accounts in small batches. Tell us what you're building; we reply within one business day.

Request early access See pricing