Project 42
Start learning
Module 3 of 3 · 45 min

DeepSeek API Workflows, Tool Calling & Cost Optimization

Build high-throughput, low-cost applications using the DeepSeek API, prompt caching, and function calling contracts.

Core concept

By the end

You will be able to

  • Integrate DeepSeek OpenAI-compatible API endpoints in TypeScript and Python.
  • Leverage automatic context caching for up to 90% input token cost reduction.
  • Build resilient tool-calling workflows with schema validation and retry policies.
01

OpenAI-Compatible API with Context Caching

DeepSeek exposes an OpenAI-compatible API, allowing direct reuse of standard SDKs by changing `base_url`. Context caching automatically identifies prefix matches across requests, reducing input token pricing significantly.

DeepSeek Client with Tool Calling
python
from openai import OpenAI

client = OpenAI(
    api_key="your-deepseek-api-key",
    base_url="https://api.deepseek.com"
)

response = client.chat.completions.create(
    model="deepseek-chat",
    messages=[{"role": "user", "content": "What is 2+2?"}],
    stream=False
)
print(response.choices[0].message.content)

Practice activity

Implement DeepSeek Tool-Calling Agent with Context Caching

  1. Create an OpenAI-compatible client targeting `https://api.deepseek.com`.
  2. Define a typed weather tool and execute function calling loop.
  3. Verify cached token hit rate across 5 sequential requests.

What to produce

  • API execution receipt showing cached token discounts and tool execution outputs.

Reflect before continuing

How does prompt structure and prefix stability affect context cache hit rates?

Evidence

Sources and verification

Knowledge check

Make it stick.

Pass at 80%

Choose the strongest answer for each question. Your attempts become part of your account transcript.

01How do you maximize context cache hit rates on the DeepSeek API?