@tokenlabai/dsh-provider
TokenLab provider bundle for DeepSeek Harness with native Responses, Messages and Chat routing plus multimodal and async tools.
162 results
TokenLab provider bundle for DeepSeek Harness with native Responses, Messages and Chat routing plus multimodal and async tools.
DeepSeek Harness plugin: route an agent request to a vision model (deepseek-v4-flash-vision-exp) when the user message contains an image, and keep the original model otherwise.
DSH 模型视觉开关:设置页「模型视觉」行,为 llm-pi-ai 自定义路由下手写声明的模型切换 input 视觉模态(settings.mutate 官方通道,热生效)。
A project-scoped multimodal video director plugin for DeepSeek Harness.
Automatic per-task model hot-switching for DeepSeek Harness (dsh): image-aware tasks route to the vision model automatically, every other task keeps your default model. Zero extra tokens, no context disturbance.
DSH 插件:提供文本 / 图像 / 视频 / 语音 / 音乐 五个生成工具,模型从 llm-pi-ai 自动发现,含 Settings UI (llm-multimodal namespace)。
Automatic reasoning and image capability detection for custom DeepSeek Harness models
DeepSeek Harness 视觉助手插件:给没有视觉能力的模型配一个可随时切换的多模态识别模型。输入框图片自动落盘并改写为文本提示,主模型调用 vision_recognize 工具即可完成看图;识别模型在 settings 的 vision-assist 命名空间热更新切换。
DeepSeek Harness web plugin: a Model Capabilities settings page for declaring model input modalities, context window, output cap and image budgets without hand-editing settings.yaml.
DeepSeek Harness plugin: analyse images out of band with a vision model — pasted images are digested into text before admission and a describe_image tool covers image paths, all without ever changing the session's model.
Give DeepSeek Harness agents eyes: local image analysis via vision models — um_analyze_img tool + model-capability recognition + hot-switch settings UI
Patient auto-retry for DeepSeek Harness: honor the upstream gateway's retry_after_seconds on a 429 capacity cooldown instead of giving up after two fast retries, and wait out the mislabeled modality 400 a saturated gateway replays from its client-error circuit. Optional floating countdown badge included.
Declare whether a configured third-party model accepts image (multimodal) input; writes the modality into the owning provider settings and verifies it through runtime model resolution.
Read-only in-conversation file preview for DeepSeek Harness: a session-authorized host Remote and a sanitizing browser modal, installed as one profile bundle.
Voice for DeepSeek Harness backed by Xiaomi MiMo: browser-native 🎤/🔊 UI (MiMo TTS read-aloud) + voice_transcribe/voice_speak calling MiMo ASR/TTS directly, with a configurable voice map and in-conversation speech strips. Fork of zhuiyueya/dsh-voice (MIT).
在 dsh 模型目录的每一行直接设置该模型的思考强度范围与支持的输入模态 (Per-model reasoning-effort range and accepted input modalities, right in the dsh Models catalog)
DeepSeek Harness visual explainer and media suite: 3Blue1Brown Manim mathematical animations, native gallery drawer, and multi-modal inline file publishing.
A headless-first macOS Computer Use driver for DeepSeek Harness with AX-first observations, window-bound multimodal evidence, host-owned approval, and evidence-backed action receipts