feat: add Xiaomi MiMo provider support

[HUQIANTAO]

Summary

• Add first-class Xiaomi MiMo provider support with dedicated API integration via token-plan endpoint (https://token-plan-cn.xiaomimimo.com/v1)
• Support two MiMo models: mimo-v2.5-pro (reasoning flagship) and mimo-v2.5 (omni)
• Implement MiMo-specific thinking toggle (thinking: { type: "enabled"/"disabled" }) instead of DeepSeek-style reasoning_effort
• Enable /models command for Xiaomi provider to fetch available models from the API
• Add provider-specific model picker UI with MiMo models and simplified thinking effort (off/high)

Changes

Provider registration:

• crates/config/src/lib.rs — ProviderKind::Xiaomi variant, default model/base_url constants, parse/as_str
• crates/tui/src/config.rs — ApiProvider::Xiaomi variant, provider_capability (thinking enabled), all match arms
• crates/cli/src/lib.rs — ProviderArg::Xiaomi, provider list, env var mapping

Model picker & validation:

• crates/tui/src/tui/model_picker.rs — Provider-specific model lists (PICKER_MODELS_XIAOMI) and thinking effort lists (PICKER_EFFORTS_MIMO with only off/high)
• crates/tui/src/config.rs — normalize_model_name_for_provider accepts MiMo model names for Xiaomi provider

API integration:

• crates/tui/src/client.rs — apply_reasoning_effort sends thinking: { type: "disabled" } for off, thinking: { type: "enabled" } for high/max
• crates/tui/src/client/chat.rs — requires_reasoning_content() includes mimo prefix for reasoning token replay
• crates/tui/src/tui/ui.rs — Allow /models command for Xiaomi provider

Model registry:

• crates/agent/src/lib.rs — ModelInfo entries for mimo-v2.5-pro and mimo-v2.5 with aliases

Config & docs:

• config.example.toml — [providers.xiaomi] section with token-plan URL and model examples
• .env.example — MIMO_API_KEY, MIMO_BASE_URL, MIMO_MODEL documentation

Test plan

• cargo fmt --all -- --check passes
• cargo clippy --workspace --all-targets --all-features --locked -- -D warnings passes
• cargo test --workspace --all-features --locked passes (4 pre-existing failures unrelated to this change)
• Manual TUI test: /provider → Xiaomi MiMo, /model → mimo-v2.5-pro, send a message
• Manual TUI test: /model → switch thinking effort off/high, verify API accepts the request
• Manual TUI test: /models → verify model list is fetched from Xiaomi API
• Manual TUI test: restart app → verify Xiaomi provider and MiMo model are persisted

Greptile Summary

This PR adds first-class Xiaomi MiMo provider support, wiring ProviderKind::Xiaomi / ApiProvider::Xiaomi through every layer of the stack — config, secrets, CLI, TUI engine, model picker, reasoning-effort translation, and reasoning-content replay. The implementation is comprehensive and follows the existing provider-addition pattern faithfully.

• API integration: apply_reasoning_effort maps MiMo's thinking: { type: \"enabled\"/\"disabled\" } format (no reasoning_effort field) across all effort levels; requires_reasoning_content and provider_accepts_reasoning_content are extended so multi-turn tool-call reasoning content is replayed correctly per MiMo docs.
• Model picker refactor: The hide_deepseek_models boolean is replaced with provider-specific provider_models/provider_efforts slices; Xiaomi gets PICKER_MODELS_XIAOMI (auto, mimo-v2.5-pro, mimo-v2.5) and a two-item effort list (off/high).
• /models support: Xiaomi is allowed to fetch available models despite living inside provider_passes_model_through, achieved via a && !is_xiaomi carve-out in the FetchModels gate.

Confidence Score: 5/5

Safe to merge — all Xiaomi variant arms are exhaustive (Rust compile-time enforcement), the reasoning-content replay path is consistent with existing providers, and manual end-to-end testing is confirmed in the test plan.

The change is a well-scoped provider addition that follows the existing pattern exactly. Every match arm across config, secrets, CLI, TUI, and model picker is complete. The reasoning-effort translation correctly maps MiMo's thinking toggle across all effort levels. The only non-trivial design choice — placing Xiaomi in provider_passes_model_through while also supporting /models — is functional and safe, if slightly surprising to future readers.

crates/tui/src/tui/ui.rs — the FetchModels guard uses a double-negative carve-out for Xiaomi that should be understood by anyone refactoring provider_passes_model_through in the future.

Important Files Changed

Filename	Overview
crates/tui/src/client.rs	Adds Xiaomi MiMo to all three effort branches (off/high/max) of apply_reasoning_effort; the low/medium branch correctly maps MiMo to thinking-enabled too. One misleading comment.
crates/tui/src/client/chat.rs	Adds mimo prefix to requires_reasoning_content and Xiaomi to provider_accepts_reasoning_content; reasoning-content replay for multi-turn tool calls is consistent and correct.
crates/tui/src/config.rs	Thorough Xiaomi variant wiring across provider_capability, model_completion_names, normalize_model_name_for_provider, env-override path, and all match arms; no missing cases detected.
crates/tui/src/tui/model_picker.rs	Replaces hide_deepseek_models boolean with provider_models/provider_efforts slices; pass-through providers get an [auto] row, Xiaomi gets PICKER_MODELS_XIAOMI and PICKER_EFFORTS_MIMO (off/high only).
crates/tui/src/tui/ui.rs	Adds Xiaomi to /models support and render provider tag; the FetchModels guard uses a double-negative because Xiaomi is both in provider_passes_model_through and also allowed to call /models — logically correct but fragile.
crates/config/src/lib.rs	Adds Xiaomi variant with DEFAULT_XIAOMI_MODEL/BASE_URL, all match arms, TOML read/write/unset/list coverage, and env-var guard/restore in tests; comprehensive and consistent.
crates/agent/src/lib.rs	Registers mimo-v2.5-pro and mimo-v2.5 with supports_reasoning: true and provider: Xiaomi; aliases are consistent with config.
crates/cli/src/lib.rs	Adds ProviderArg::Xiaomi, PROVIDER_LIST entry, provider_slot, provider_env_vars (MIMO_API_KEY), and TUI guard for Xiaomi; all arms exhaustive.
crates/secrets/src/lib.rs	Maps xiaomi/mimo/xiaomi-mimo slug to MIMO_API_KEY; consistent with all other provider key lookups.

Sequence Diagram

sequenceDiagram
    participant User
    participant TUI
    participant Config
    participant Client
    participant MiMoAPI

    User->>TUI: /provider → Xiaomi, /model → mimo-v2.5-pro
    TUI->>Config: "api_provider() == Xiaomi"
    Config-->>TUI: DEFAULT_XIAOMI_BASE_URL, MIMO_API_KEY

    User->>TUI: "Send message (effort = high)"
    TUI->>Client: apply_reasoning_effort(body, high, Xiaomi)
    Client->>Client: "body[thinking] = {type:enabled}"
    Note over Client: No reasoning_effort field (MiMo-specific)

    TUI->>Client: build_chat_request(messages)
    Client->>Client: requires_reasoning_content(mimo-v2.5-pro) → true
    Client->>Client: "should_replay_reasoning_content → replay if effort != off"

    Client->>MiMoAPI: POST /v1/chat/completions
    Note over Client,MiMoAPI: Bearer MIMO_API_KEY
    MiMoAPI-->>Client: SSE stream with reasoning_content + content
    Client-->>TUI: parsed MessageResponse
    TUI-->>User: display thinking + answer

Loading

_{Reviews (4): Last reviewed commit: "chore: bump version to 0.8.47 (dev)" | Re-trigger Greptile}

[gemini-code-assist]

Code Review

This pull request adds support for the Xiaomi MiMo Token Plan provider, integrating its models (mimo-v2.5-pro, mimo-v2.5, and mimo-v2-flash) into the configuration, CLI, and interactive TUI. The review feedback highlights three key issues: a regression in the model picker that removes the "auto" option for other pass-through providers, an issue where switching to Xiaomi with a "Max" reasoning effort silently disables thinking, and an overly restrictive model name normalization bypass that prevents the use of "auto" or custom models.

[gemini-code-assist]

For other pass-through providers, setting provider_models to &[] removes the "auto" ("select per turn") option from the model picker, which is a regression from the previous behavior where vec!["auto"] was shown. We should return &[("auto", "select per turn")] for other pass-through providers to preserve this option.

Suggested change

	let provider_models: &'static [(&'static str, &'static str)] = match app.api_provider {
	crate::config::ApiProvider::Xiaomi => PICKER_MODELS_XIAOMI,
	p if crate::config::provider_passes_model_through(p) => &[],
	_ => PICKER_MODELS,
	};
	let provider_models: &'static [(&'static str, &'static str)] = match app.api_provider {
	crate::config::ApiProvider::Xiaomi => PICKER_MODELS_XIAOMI,
	p if crate::config::provider_passes_model_through(p) => &[("auto", "select per turn")],
	_ => PICKER_MODELS,
	};

[gemini-code-assist]

If the user has ReasoningEffort::Max selected (which is common for DeepSeek), switching to Xiaomi MiMo will result in normalized being Max. Since PICKER_EFFORTS_MIMO does not contain Max, the position search will fail and default to 0 (ReasoningEffort::Off), silently disabling thinking. We should map ReasoningEffort::Max to ReasoningEffort::High when is_xiaomi is true.

        let selected_effort_idx = provider_efforts
            .iter()
            .position(|e| {
                if is_xiaomi && normalized == ReasoningEffort::Max {
                    *e == ReasoningEffort::High
                } else {
                    *e == normalized
                }
            })
            .unwrap_or(0);

[gemini-code-assist]

Restricting the bypass to only models starting with "mimo" prevents the use of "auto" (the special TUI-level model name for auto-selection) or any custom/fine-tuned models that do not follow the "mimo" prefix. Since Xiaomi is a pass-through provider, we should allow any valid model name format.

    if matches!(provider, ApiProvider::Xiaomi) {
        if trimmed
            .chars()
            .all(|ch| ch.is_ascii_alphanumeric() || matches!(ch, '-' | '_' | '.' | ':'))
        { 
            return Some(trimmed.to_string());
        }
        return None;
    }