Add Qiniu Cloud (七牛云) as a built-in model provider

[HannahKking]

Summary

This PR adds Qiniu Cloud AI Inference Hub (七牛云) as a built-in model provider.

Qiniu Cloud provides an OpenAI-compatible API at https://api.qnaigc.com/v1 with access to 50+ mainstream large models, making it straightforward to integrate as a native Plandex provider.

Changes

• app/shared/ai_models_providers.go

  • Add ModelProviderQiniu constant ("qiniu")
  • Add QiniuApiKeyEnvVar constant ("QINIU_API_KEY")
  • Register ModelProviderQiniu in AllModelProviders
  • Add provider config in BuiltInModelProviderConfigs with base URL https://api.qnaigc.com/v1

• app/shared/ai_models_available.go

  • Add Qiniu as an additional provider for deepseek/v3 (model ID: deepseek-v3) and deepseek/r1 (model ID: deepseek-r1)
  • Add qiniu/kimi-k2 model — Moonshot AI Kimi K2.5, 256K context window
  • Add qiniu/glm-5 model — Zhipu AI GLM-5, 128K context window

Setup

Users need to set one environment variable:

export QINIU_API_KEY=<your_api_key>

API keys can be obtained at https://portal.qiniu.com/ai-inference/api-key. New users receive 3 million free tokens.

No client.go changes needed

Qiniu's API is fully OpenAI-compatible, so the existing newClient() generic path handles it without any provider-specific logic.

References

• API docs: https://developer.qiniu.com/aitokenapi/12882/ai-inference-api
• Chat completions: https://developer.qiniu.com/aitokenapi/13211/chat_completions
• Model list: https://developer.qiniu.com/aitokenapi (query /v1/models for full list)

Test plan

• Set QINIU_API_KEY and verify Qiniu provider appears in plandex models available
• Test deepseek/v3 via Qiniu provider
• Test deepseek/r1 via Qiniu provider
• Test qiniu/kimi-k2 model
• Test qiniu/glm-5 model

🤖 Generated with Claude Code