Skip to content

Providers

Each provider has its own Behavior page because the web UIs don't all expose the same controls, and some of them have unique quirks that need extra explanation. If you are looking for a specific toggle, model picker, or provider quirk, this is the place to find it.

App-wide settings live elsewhere

Accounts, sessions, formatting, ports, API keys, and logging are covered under Features. This section is for provider-specific controls like Thinking, Search, model pickers, uploads, blocked-response handling, and provider quirks.

  • DeepSeek DeepSeek Behavior

    DeepThink, Search, file uploads, blocked-response handling, and chat reuse.

    ➡ Open DeepSeek

  • Z.ai GLM Behavior

    Deep Think, Search, Advanced Search, Tools, model selection, token counting, uploads, and GLM-specific caveats.

    ➡ Open GLM

  • MoonshotAI Kimi Behavior

    Thinking, Search, file uploads, blocked-response handling, and Kimi account-tier notes.

    ➡ Open Kimi

  • Qwen QwenLM Behavior

    Thinking, Web Search, model selection, token counting, uploads, and provider guardrails.

    ➡ Open QwenLM

  • Perplexity Perplexity Behavior

    Model selection, Thinking, Search, uploads, and email-code login notes.

    ➡ Open Perplexity

  • HuggingFace HuggingChat Behavior

    Model selection, inference provider, Thinking Effort, Exa search, uploads, and account limits.

    ➡ Open HuggingChat

  • Google AI Studio Google AI Studio Behavior

    Gemini model selection, Thinking Level, Search grounding, URL Context, uploads, sampling, and humanized mouse movement.

    ➡ Open AI Studio

  • Xiaomi Xiaomi MiMo Behavior

    MiMo model selection, thinking output filtering, uploads, regional availability notes, token usage, and chat reuse.

    ➡ Open MiMo

Quick Picker

Provider Best doc to read when...
DeepSeek You need DeepThink/Search controls, duplicate-prompt reuse, or the DeepSeek refusal-message workaround.
GLM Chat You need Tools, token counting, model picker behavior, Repetition Buster, or timing/instability notes.
Kimi / Moonshot You use Kimi's Thinking/Search modes, Google login, uploads, or want to understand its account-tier caveats.
QwenLM You need Qwen Thinking, Web Search, model selection, token counting, file uploads, or guardrail behavior.
Perplexity You use Perplexity's model picker, Thinking/Search behavior, uploads, or email-code login.
HuggingChat You need model/provider selection, Thinking Effort, Exa search, uploads, or help with monthly credit limits.
Google AI Studio You use Gemini through AI Studio, need model/Thinking/Search controls, etc.
Xiaomi MiMo You need MiMo model selection, thinking output filtering, proxy setup, or regional availability notes.

Provider docs are intentionally a little more detailed than the overview pages, though they're still focused on practical user-facing info, as I wanted to avoid dumping too much technicalities into the main flow. If you want to understand more about how a provider integration works under the hood, you're going to have to dig through the code or ask in the community.