Alice-Marx/dsh-model-router

Model Router and GALGame plugin for DeepSeek Harness

What it solves

This project provides a cost-aware workbench for DeepSeek Harness Desktop (DSH) that helps users plan tasks, route them to specific AI models or official tools, and manage execution permissions. It addresses the need for structured task delegation, budget tracking, and dependency-aware workflows within the DSH ecosystem.

How it works

The workbench operates through a four-tab interface that allows you to describe tasks, configure model routes, manage budgets, and interact with official tools. You can choose between automated routing recommendations or manual assignments. The system separates API spending from subscription usage, provides live progress tracking, and supports both read-only execution and editable CLI teamwork within either a restricted sandbox or a native environment.

Who it’s for

It is designed for users of DeepSeek Harness Desktop who need to manage complex, multi-step AI tasks that require specific model selection, cost monitoring, and controlled execution environments.

Highlights

  • Cost-Aware Planning: Estimates token costs and manages daily/monthly budget reservations before executing tasks.
  • Flexible Execution: Supports read-only model API calls and interactive CLI teamwork with explicit permission controls.
  • Dependency Management: Decomposes complex requests into dependent work packages that can be routed to different models or tools.
  • Execution Transparency: Provides live elapsed time, progress tracking, and detailed history of results, costs, and failures.
  • Safety Controls: Offers a choice between a restricted Harness sandbox and a native environment for CLI operations.

Related

  • Project
    professorpalmer/Puppetmaster

    A supervisor and job store for AI agent CLIs and provider APIs that ensures durable state and reviewable output for complex engineering tasks using a SQLite backend.

  • Project
    anomalyco/browser-control

    Browser Control is an open‑source Node.js driver that lets AI agents or CLI users run Playwright code inside their own Chromium browser (with all logged‑in sessions and extensions). It works via a local relay and a browser extension, supports persistent sessions, inspection helpers, human hand‑offs, video/ HAR recording, and a TypeScript Effect client for secret‑safe API calls. Intended for trusted local use only.

  • Project
    Anil-matcha/Open-AI-Design-Agent

    Open AI Design Agent is an MIT‑licensed, self‑hostable AI design assistant that turns plain‑language briefs into complete visual campaigns (posters, logos, social graphics, video ads, etc.). It runs an autonomous planning loop that selects from 200+ image and video models via the Muapi.ai API, supports brand‑kit conditioning, multi‑image reference, edit mode, templates, and a visual workflow editor. Install with Node.js + FastAPI or Docker, or use the hosted web UI.

  • Project
    VisActor/skills

    VisActor/skills provides three plug‑in “skills” (vchart‑skill, vtable‑skill, data‑storytelling) that let an LLM‑backed agent generate interactive charts, sortable tables, and HTML‑based data stories directly from CSV or similar files. Installation is a one‑line `npx skills add …` command; the skills run locally using Node, Python, and CDN‑served JS libraries, requiring no separate VisActor server or API key.