Skip to content

Feat/deepseek provider - #430

Open
NOLANeight wants to merge 1 commit into
bytedance:mainfrom
NOLANeight:feat/deepseek-provider
Open

Feat/deepseek provider#430
NOLANeight wants to merge 1 commit into
bytedance:mainfrom
NOLANeight:feat/deepseek-provider

Conversation

@NOLANeight

@NOLANeight NOLANeight commented Jul 23, 2026

Copy link
Copy Markdown

Closes #406

Add a new DeepSeek client that uses the standard Chat Completions API
via the OpenAICompatibleClient base, matching the pattern used by the
existing Doubao provider.

Changes:

  • New deepseek_client.py: DeepSeekClient + DeepSeekProvider
  • Register DEEPSEEK in LLMProvider enum and LLMClient factory
  • 12 unit tests covering init, message handling, tool-calling, and config
  • Update README with DeepSeek examples and env var docs

@CLAassistant

CLAassistant commented Jul 23, 2026

Copy link
Copy Markdown

CLA assistant check
All committers have signed the CLA.

@NOLANeight
NOLANeight force-pushed the feat/deepseek-provider branch from 21e02cc to a10a8ae Compare July 23, 2026 07:16
Add a new DeepSeek client that uses the standard Chat Completions API
(chat.completions.create) via the OpenAICompatibleClient base, matching
the pattern used by the existing Doubao provider.

Changes:
- New deepseek_client.py: DeepSeekClient + DeepSeekProvider (82 lines)
- Register DEEPSEEK in LLMProvider enum and LLMClient factory
- Add 12 unit tests covering provider config, client init, message
  handling, tool-calling support, and YAML config integration
- Update README with DeepSeek examples and env var docs

Closes bytedance#406
@NOLANeight
NOLANeight force-pushed the feat/deepseek-provider branch from a10a8ae to 4861535 Compare July 23, 2026 07:25
@NOLANeight

Copy link
Copy Markdown
Author

@cla-bot check

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Add DeepSeek V4 Flash & Pro with reasoning/non-reasoning modes

2 participants