Articles tagged "llm-pi-ai"

dsh-llm-model-tuning：在 Models 页按模型配置推理强度与多模态输入

2026-10-21 DeepSeek Harness plugin DeepSeek DSH dsh-llm-model-tuning model inference llm-pi-ai

dsh-llm-model-tuning 是 xurunxin 维护的 DSH 客户端 bundle，在 Settings → Models 的自定义 provider 卡片上为每个模型编辑 reasoningEfforts、input 及路由级 defaultInput，写入 llm-pi-ai 段且无需重启。内置编辑器 deliberately 不提供 provider 级 reasoning 控件；本插件以 per-model 粒度避免同一路由下模型能力不一致导致的拒绝。要求 dsh 0.1.7+（configForms）且 profile 已 compose llm-pi-ai。安装使用 plugin_manager install_bundle 指向包目录绝对路径，symlink 安装支持改源码后 HMR 或 profile 重启。

dsh-llm-pi-ai-live：为 llm-pi-ai 路由实时刷新模型列表

2026-10-21 DeepSeek Harness plugin DeepSeek DSH dsh-llm-pi-ai-live Networking tools llm-pi-ai

DSH 的 dsh-llm-pi-ai 适配器默认从 pi-ai 静态快照展示模型，新模型与手写列表易陈旧。dsh-llm-pi-ai-live 通过官方 settings 接缝向各路由 live 端点拉取模型 id，合并 catalog 元数据并以追加方式写回配置；支持定时、启动与 refresh_model_catalog 工具，失败不改原配置。文章介绍安装（npm build + dsh plugin add link）、dryRun 首次验证及主要配置项，并说明适用场景与安全注意。

dsh-model-ultra：为 DSH 0.2.0 提供 llm-pi-ai 提供商全生命周期管理

2026-10-21 DeepSeek Harness plugin DeepSeek DSH dsh-model-ultra model inference llm-pi-ai

dsh-model-ultra 是 MrCashmere 维护的 DSH 静态 bundle 插件，兼容 >=0.2.0-rc.1，在设置页提供「模型 Ultra」界面，覆盖 llm-pi-ai 提供商 CRUD、加密密钥、远端模型发现、逐模型思考强度、OpenRouter 路由、智能路由与组合、探活观测及连通性测试，卸载时可还原 providers。本文基于官方 README 介绍安装命令、典型步骤与升级注意事项。

dsh-model-sync：把推理端模型列表自动写进 DSH 设置

2026-10-21 DeepSeek Harness plugin DeepSeek DSH dsh-model-sync model inference llm-pi-ai

dsh-model-sync 是 mpetruc 维护的 DSH 模型推理类插件（v0.7.1，MIT）。它通过 llm 服务的 discoverModels，把各 llm-pi-ai provider 公布的模型镜像到 Settings → Models，减少手填 model id。激活或 Plugins 页重开 model-sync 约 1.5 s 后同步；可配置 intervalMinutes 与 prune，并用 owner 标记区分手填与自动项。同步摘要写入 lastSync 并在插件卡片展示；rejected 状态存于 storages/model-sync/state.json。endpoint 异常时跳过并告警，不拖垮整插件。

hytime/dsh-thinking-effort: Supplementing the Thinking Intensity Settings for DSH's llm-pi-ai Model

2026-08-29 DeepSeek Harness plugin DeepSeek plugin dsh plugin llm-pi-ai Inference Strength dsh-thinking-effort

`@hytime/dsh-thinking-effort` is a DSH plugin designed to supplement default configurations for third-party `llm-pi-ai` models that lack declared thinking effort levels. Its core functionalities include adding default levels of `off`, `high`, and `max` without overriding custom values; supporting the selection of thinking effort levels per model within DSH settings and mapping DSH levels to corresponding gateway values (e.g., `high` mapping to `ultra`); setting default thinking intensity for sub-agents, which only applies when not explicitly specified in a request; supporting multilingual interfaces in Chinese, English, Japanese, and Korean (with Japanese/Korean requiring DSH core support); and displaying a version watermark on the settings page. Users can manage the plugin via the DSH CLI command `dsh plugin --pr

DeepSeek Harness Model Configuration Plugin: Managing Custom Endpoints and Model Parameters in Web UI

2026-08-26 DeepSeek Harness plugin DeepSeek DSH DeepSeek Harness plugin Model Configuration llm-pi-ai

marvekg/deepseek-harness-model-config adds a "Model Advanced Configuration" page to the DeepSeek Harness Web UI, supporting the creation of OpenAI/Anthropic protocol endpoints, pulling or manually adding models, and populating context and inference parameters via models.dev. API keys are stored through Harness credentials and are not written to settings.yaml. Installation: dsh plugin --profile web add github:MarvekG/deepseek-harness-model-config, usable under "Settings → Model Advanced Configuration." Licensed under MIT, maintained by MarvekG.