from_provider doesn't support Litellm as described in documentation

[kouk]

• This is actually a bug report.
• I am not getting good LLM Results
• I have tried asking for help in the community on discord or discussions and have not received a response.
• I have tried searching the documentation and have not found an answer.

What Model are you using?

• gpt-3.5-turbo
• gpt-4-turbo
• gpt-4
• Other (please specify)

Describe the bug

In the documentation it is shown that to use the litellm provider you can call instructor.from_provider. However that method doesn't support litellm.

To Reproduce

Run the simple example from the documentation. You will get:

Traceback (most recent call last):
  File ".../test.py", line 6, in <module>
    client = instructor.from_provider("litellm/gpt-3.5-turbo")
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".../.venv/lib/python3.12/site-packages/instructor/auto_client.py", line 475, in from_provider
    raise ConfigurationError(
instructor.exceptions.ConfigurationError: Unsupported provider: litellm. Supported providers are: ['openai', 'azure_openai', 'anthropic', 'google', 'mistral', 'cohere', 'perplexity', 'groq', 'writer', 'bedrock', 'cerebras', 'fireworks', 'vertexai', 'generative-ai', 'ollama', 'xai']

Expected behavior

It works.

Screenshots
If applicable, add screenshots to help explain your problem.

[jxnl]

@claude can you take a look and make sure we have the correct routing in auto_client to support litellm then make a PR

[claude]

Claude finished @jxnl's task —— View job • claude/issue-1710-20250715_191904 • Create PR ➔

---

✅ LiteLLM support added to from_provider!

I've successfully fixed the missing LiteLLM support in the from_provider function.

Todo List:

• Read CLAUDE.md for repo guidelines
• Examine current auto_client.py implementation
• Check for existing LiteLLM integration in codebase
• Add LiteLLM support to from_provider function
• Test the fix with the example from the issue
• Run linting and tests per repo guidelines
• Commit and push changes
• Create PR

Changes Made:

✅ Fixed the ConfigurationError - The example from the issue documentation now works:

client = instructor.from_provider("litellm/gpt-3.5-turbo")
async_client = instructor.from_provider("litellm/gpt-3.5-turbo", async_client=True)

✅ Added "litellm" to supported providers list (line 34)

✅ Implemented complete litellm provider handler (lines 855-886) that:

• Imports completion and acompletion from litellm package
• Handles both sync and async clients appropriately
• Uses the existing from_litellm function
• Sets proper default mode (instructor.Mode.TOOLS)
• Includes comprehensive error handling for missing packages
• Follows the same pattern as other provider implementations

Commit Details:

• Commit: 3b70ca5 - fix(auto_client): add support for litellm provider in from_provider
• Branch: claude/issue-1710-20250715_191904
• Files: 1 changed, 34 insertions

The fix resolves the discrepancy between the documentation and implementation, allowing users to use LiteLLM through the convenient from_provider interface as documented.

---