feat: add Novita AI as an LLM provider

[Alex-wuhu]

Summary

• Adds Novita AI to the backend PROVIDER_REGISTRY in llm_client.py as an OpenAI-compatible provider
• Endpoint: https://api.novita.ai/openai
• Default max tokens: 8192
• Adds corresponding entry to the frontend FALLBACK_LLM_PROVIDERS list in EnterpriseSettings.tsx

Models

Suggested models to configure after adding the provider:

• moonshotai/kimi-k2.5 (default)
• deepseek/deepseek-v3.2
• zai-org/glm-5

Test plan

• Add a Novita AI entry in Enterprise Settings → LLM Pool with a valid NOVITA_API_KEY
• Verify the provider appears in the provider dropdown
• Test a chat completion request through an agent configured with the Novita provider

🤖 Generated with Claude Code

[wisdomqin]

Hi @Alex-wuhu, thanks for the contribution!

We really appreciate you taking the time to add Novita AI support. However, to keep our provider list focused and manageable, we are currently prioritizing only the most widely-used mainstream foundation model providers (like OpenAI, Anthropic, Google, DeepSeek, etc.) and major local inference engines.

For other OpenAI-compatible platforms like Novita, users can easily connect to them by selecting the Custom provider option in Enterprise Settings and pasting the https://api.novita.ai/openai base URL.

Because of this, we'll decline this PR for now to keep the default provider list concise. Thank you again for your understanding and your interest in contributing to Clawith!