Models & Providers

Declares image input on relay routes whose endpoint omits modality metadata, so models that can see photos stop being rejected as text-only. This also overrides routes that are explicitly declared text-only.

Multi-model deliberation pipeline for DeepSeek Harness: independent proposals, blind peer review, program-executed evidence verification, immutable revision and independent acceptance on a fresh context.

Configure DMXAPI Chat Completions, Responses, Anthropic Messages, and custom providers; edit models on existing preset routes.

Adds an independent OpenCode Go connection with credential settings, manual model discovery, and per-conversation session headers.

Registers OpenCode Zen free models as a native provider: the catalog is discovered live from opencode.ai intersected with models.dev pricing, and a detail page can re-fetch it or test which models answer. Declares no dependencies, so the host's pi-ai stays the only one.

Recognizes a model ID in the DeepSeek Harness model catalog and fills that row's context window, input modalities and reasoning levels as editable suggestions, from a bundled offline rule table covering 33 vendors.

Gives pi-ai routed models a reasoning-effort ladder without manual settings: falls back declared to models.dev to a 7-tier default, and replaces the composer model chip with a Codex-style vertical picker.

Configure thinking toggles and reasoning efforts per model for llm-pi-ai and official DeepSeek, and specify input modalities (input or inputModalities) in the DSH settings panel.

Brings the providers you already configured in cc-switch into DeepSeek Harness: reads the cc-switch database, converts Claude / Codex / OpenClaw / MCode / Pi entries into llm-pi-ai routes with their API keys, and adds a health check that measures time-to-first-byte and whether each stream completes.

Shows a Spark Scope card at the bottom of the dsh sidebar: model server decode and prefill rates with sparklines, running and queued requests, KV cache, TTFT and TPOT p95, and GPU temperature, power, load and memory for each DGX Spark node. Read-only; needs a running Spark Scope server. Works in dsh web and the desktop app.

Runs MoE4All models locally in DSH with automatic engine installation, model discovery and downloads, inference settings, and live speed monitoring, or connects to an existing service.

Adds Kimi models to native DSH sessions with per-model selection of the embedded Kimi core or DSH native execution.

Model manager for the composer picker: probes a model on demand and keeps the timestamped reason it failed, counts usage from committed assistant messages to rank models by proficiency level, pins, favourites and hides rows, and switches along a failover chain when a turn fails, asking the new model to pick the work back up.

Replaces the composer model seat with a search-first picker and adds a model management page: create and edit llm-pi-ai channels, choose which of an endpoint's models exist, and declare per-model thinking ladders and capabilities.

Reuses the local model proxy of a logged-in XiaoyiWork desktop app (Windows only) so its models appear in DeepSeek Harness without a separate API key.

Adds a thinking-depth slider mapped to declared model levels, configuration guidance, and optional inline model settings.

Connects Qoder CN desktop app models to DeepSeek Harness with zero user interaction: decrypts local OSCrypt credentials, signs requests with the COSY scheme against the CN gateway, and lists 14 models with free-tier badges.

Adds a one-click switch to every provider row in Settings - Models: turn a provider off and it leaves the model picker, turn it back on and it returns. It works by wrapping the model catalog call, so it affects the catalog and the picker only, never routing, keys or provider config; existing sessions that already use such a model keep working.

Per-model enable/disable and renaming for every provider in DeepSeek Harness: a panel at the bottom of Settings - Models with a search box, status chips with counts, a provider filter, collapsible provider groups with group-level switches, and batch actions scoped to the current view. Disabled models leave the model picker and custom names replace the built-in ones. Preferences persist in the profile and apply immediately, without reloading the page or restarting DSH.

DeepSeek Harness 的模型fallback链插件：以更智能、更人性化的方式配置fallback链。当主模型撞限流、额度耗尽，或者5xx超时的时候自动向下改派模型，回合不中断。设置页可视化编排：拖拽排序、按计费或者上下文窗口排序候选模型、失败模型自动冷却、管理多条链的停用和启用。零依赖、无构建步骤、默认不介入任何请求。

Per-model input type and thinking level declarations on the 模型 settings page — 输入类型 checkboxes on every provider card, and a 思考等级 field in the provider 编辑 dialog. Writes the official pi-ai input and reasoningEfforts fields.

OpenCode CLI inside DeepSeek Harness: an `opencode` delegation tool, a /opencode command, and an `opencode` LLM provider route — all over the ctx.subprocess seam, with `opencode models` discovery.

Installable wrapper for the DSH plugin in dsh-kdczyz-mode/: WorkBuddy model pool + Trae connector + Qoder account pool as three independent providers, plus one settings page with six tabs (XD Pool / Trae config [provider switch, account binding, model sel