The DeepSeek API lets developers call DeepSeek's language models from code using the same request formats as OpenAI's or Anthropic's APIs.
In plain words
Because the format matches, you can reuse the OpenAI or Anthropic software kits: you change the base address, supply a DeepSeek API key and pick a DeepSeek model name. That also lets coding assistants and agent tools that support those formats use DeepSeek models as their backend. A JSON output mode helps apps get replies they can parse. Old model names can be retired and redirected, so check the docs.
A quick example
Imagine a developer with an app built on the OpenAI Python library. To test DeepSeek, they change the base address to DeepSeek's, set a DeepSeek key and model name, and run their existing code unchanged otherwise.
Where you'll meet it
- Apps and agents that want DeepSeek models behind a familiar API.
- Coding assistants configured to use DeepSeek as the model.
Related
Sources
- DocsYour First API Call, DeepSeek API Docs · read 29 Sept 2026
- DocsJSON Output, DeepSeek API Docs · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.