Put the world's best AI agents to work. One workspace for all of them -- Claude Code, Codex, Gemini, Grok and more, running side by side on your Desktop.
-
Updated
Sep 22, 2026 - JavaScript
#
agent-workbench
Here are 44 public repositories matching this topic...
基于 DeepSeek Harness（DSH）的稳定桌面端，集成Git、内置浏览器与记忆功能 | DeepSeek Harness desktop GUI with local workspaces, Git, browser and memory.
agent gui ai desktop dsh ai-agents ai-agent llm deepseek vibe-coding agent-workbench deepseek-harness dsh-plugin dsh-plugins dsh-plugin-desktop
-
Updated
Nov 5, 2026 - TypeScript
TypeScript agent workbench for Web, Electron and Expo with pluggable execution hosts and managed Web updates
-
Updated
Nov 14, 2026 - TypeScript
Nomi·诺米 — 让每一种 AI 能力，都为你所用。One place. Every AI capability. Agent workbench for Claude Code, Codex, Gemini CLI and more: run them side by side, one board to watch and manage, connect a model once for every AI. Go + Wails, ~25 MB; optional desktop character.
mcp codex ai-agents mascot wails local-first nomi desktop-ai coding-agents gemini-cli llm-gateway claude-code agent-workbench
-
Updated
Nov 14, 2026 - TypeScript
Open-source multi-model AI agent workbench (Desktop GUI, TUI & Web) — dispatch subagents across Claude, GPT, Gemini & Kimi, bridge WeChat/QQ, and run locally.
desktop-app cli mcp multi-agent developer-tools codex tauri kimi terminal-ui local-first ai-agent llm claude-code coding-agent agent-workbench
-
Updated
Nov 13, 2026 - TypeScript
Local-first AI agent workbench for long-running projects: background work, built-in memory (no Docker), six web search services, MCP. macOS, Linux, Windows.
electron desktop-app cli mcp openai glm claude web-search long-term-memory local-first ai-agent brave-search llm deepseek firecrawl self-evolving model-context-protocol openviking proactive-agent agent-workbench
-
Updated
Oct 23, 2026 - JavaScript
FF - FDE 交付工作台完整私有交接快照：源码、数据分析项目、阶段状态与历史会话
-
Updated
Nov 13, 2026 - TypeScript
My OpenCode workbench setup and agent workflow patterns.
-
Updated
Sep 8, 2026 - Shell
现代化 DeepSeek Harness (DSH) 全栈增强工作台 · 多源 MCP 可视化调试中枢 · 视觉思考引擎 · 灵动极客伴侣
deepseek model-context-protocol mcp-manager agent-workbench deepseek-harness dsh-plugin dsh-plugin-market dsh-better-sidebar
-
Updated
Aug 16, 2026 - TypeScript
Agent workbench for auto research, self-improvement loops, and long-running work — multi-agent teams, persistent project memory, cross-machine GPU training. Developed entirely inside Cortex.
gpu multi-agent autonomous-agents ai-agents ml-training recursive-self-improvement coding-agents claude-code auto-research agent-workbench
-
Updated
Nov 13, 2026 - TypeScript
A workbench for you and your agents.
-
Updated
Nov 14, 2026 - TypeScript
本地多运行时 Agent 工作台：围绕材料研究、生成可编辑报告、独立复核、追溯证据，并从人工反馈持续学习。
workflow business audit provenance human-in-the-loop ai-agents evidence briefing report-generation local-first multi-runtime research-agent agent-workbench
-
Updated
Oct 17, 2026 - JavaScript
One local-first desktop workspace for every AI coding agent — Codex, Claude, OpenCode and your own models.
electron desktop-app open-source typescript mcp opencode multi-model codex ai-agents claude local-first ai-agent llm ai-coding claude-code coding-agent ai-coding-workflow agent-workbench
-
Updated
Nov 13, 2026 - TypeScript
A powerful, lightweight, cross-platform AI workspace supporting local GGUFs and multi-provider APIs.
desktop-app rust cross-platform mcp tauri solidjs ai-assistant ai-tools ai-agent llm chatgpt-client llama-cpp local-llm agent-workbench
-
Updated
Oct 30, 2026 - Rust
Local-first team workbench for AI agents: project mailboxes, shared context, task handoffs and human review. v0.8.2 · macOS App and local Web; Intel, Windows and Linux previews.
python json mcp multi-agent mailbox stdio codex ai-agents agent-management local-first claude-code agent-workbench ai-agent-collaboration
-
Updated
Oct 31, 2026 - Python
Java-first open-source AI Agent framework & workbench. Core framework: general-purpose ReAct agents and harness agents (sandbox, HITL approval, resumable execution). AgentForge Service: best-practice standard agent infrastructure to build, debug, publish and operate production-ready agents — streaming, ChatMemory, MCP/HTTP tools, session replay.
java machine-learning streaming ai mcp artificial-intelligence openai java-agent ai-agents agent-framework ai-agent llm anthropic react-agent agentic-ai tool-calling mcp-client agentforge agent-workbench
-
Updated
Nov 13, 2026 - Java
Claude 白白板 - 榨干每一单位 token 效率的 Claude Code 多 Agent 白板工作台 | Maximize Every Token
nodejs javascript automation typescript wiki multi-agent kanban task-scheduler claude skill-management ai-agent claude-code agent-workbench
-
Updated
Apr 21, 2026 - JavaScript
Open-source graphical Agent workbench for long-running coding tasks, with controlled edits, permissions, checkpoints, and event-backed recovery.
typescript event-sourcing self-hosted developer-tools local-first crash-recovery ai-agent llm ai-coding coding-agent agent-workbench
-
Updated
Sep 3, 2026 - TypeScript
个人 Agent 工作环境：Governance Kernel / Overlay / Project / Git / Harness 之上的只读 Verified Projection GUI（Electron + React + TypeScript）。投影层不替代 Harness、不复制外部正本。
-
Updated
Sep 17, 2026 - TypeScript
基于 Claude Code、Pi Agent 和 Herdr 的多 Agent 协作开发工作流，涵盖任务分级、模型路由、Skills、独立审查与开发者验收
-
Updated
Nov 12, 2026 - JavaScript
Add this topic to your repo
To associate your repository with the agent-workbench topic, visit your repo's landing page and select "manage topics."