evoflux
Discovered★ 9Evoflux is an open-source, local-first workspace where AI agents build software, conduct deep research, automate browser tasks, and collaborate in parallel. Connect any model, keep control of your wor
EvoFlux
A local-first desktop workspace for AI agent teams.
EvoFlux brings agent conversations, files, terminal, browser, git, and verification into one workspace. Use Work for general tasks and Coding for repository-aware development.
Highlights
- Work and Coding modes for cowork tasks and persistent repositories.
- Lead-and-specialist teams with per-agent models, skills, tools, and permissions.
- Bring your own model through hosted, cloud, routed, or local providers.
- Repository-aware Coding with code search, symbol relationships, LSP feedback, git workflows, and reviewable diffs.
- Spec-first delivery through Evo Agent Specs (EASD), with user-approved specifications, verification, and evidence-based convergence.
- Local control with inspectable history, permissions, sandboxing, and outbound data protection.
- Memory, plugins, MCP, scheduling, and browser tooling are available as integrated workspace capabilities.
Download
Download the latest desktop release from GitHub Releases. Packages are available for macOS, Windows, and Linux. The desktop packages include the native Python sidecar.
Linux packages should be installed through the system package manager, for example:
sudo apt install ./EvoFlux_*_amd64.deb
WebBridge is an optional browser companion distributed separately in the evo-webbridge repository.
Run from source
Requirements
Setup
git clone https://github.com/evoelsewhere/evoflux.git
cd evoflux
uv sync
cd web && bun install && cd ..
Start the development app with:
# React UI + local API
make dev-web
# React UI + local API + Tauri desktop shell
make dev-desktop
make dev is an alias for make dev-web. EvoFlux is developed and shipped as
a desktop product; the web development server is an internal development
surface.
On first launch, connect a model provider, then start a Work session or open a repository in Coding mode.
Documentation
The README stays intentionally high-level. Detailed contracts and guides live
under documents/:
- Documentation index
- EASD methodology
- Evo Agent Specs guide
- Agent Plugins guide
- Coding architecture
- Development guide
Architecture at a glance
Tauri desktop shell
↓
React UI ↔ local FastAPI sidecar
↓
local state, agent runtime, repositories, and model providers
The app keeps the main runtime and project state on the user's machine. Model providers are replaceable, while the harness handles context, tools, permissions, memory, and verification.
Project layout
app/ FastAPI sidecar, agents, code context, memory, scheduler, MCP
web/ React interface
desktop/ Tauri shell and Python sidecar packaging
seed/ Work and Coding blueprints, skills, and config
tests/ Backend and frontend tests
documents/ Architecture, feature contracts, guides, and records
Contributing
Issues and pull requests are welcome. Keep changes focused and include the smallest relevant test run. Report vulnerabilities privately through GitHub Security Advisories.
License
EvoFlux is released under the Apache License 2.0.
Similar plugins
by Devin-AXIS
Enterprise-grade, local-first Agent Workbench for people and agent teams. A unified multi-engine workspace for Codex Harness, DeepSeek Harness, and OpenCode, with unified plugins and Skills, multi-age
★ 6.1k
↓ 459/wk
NOASSERTION
TypeScript
Sep 16, 2026
dsh plugin --profile web add deepseek-idesignby mjylfz
Adds a subagent_codex tool that delegates tasks to the local OpenAI Codex CLI, running the work in an isolated external process.
★ 4
MIT
JavaScript
Aug 25, 2026
dsh plugin --profile web add dsh-subagent-codexby TNJ2026
Local-first, durable workflow runtime for Agent Apps—compile static Workflow DSL to LangGraph and orchestrate trusted multi-agent execution through Codex, WorkBuddy, DeepSeek Harness, or any MCP clien
★ 3
Python
Sep 11, 2026
dsh plugin --profile web add promptaflow-ui-testsby alib8b8
Local-first automation Agent & deterministic workflow engine — the secure control layer between AI and your data. Data stays local. Connect your own LLM / files / notes / databases via Connector API.
★ 14
↓ 41/wk
NOASSERTION
Go
Sep 8, 2026
dsh plugin --profile web add @alib8b8/dsh-plugin-aflareby GZX2211
专为 DeepSeek Harness Web GUI 打造的可视化多 Agent 工作流编排插件。公开测试版已上线!接下来会随着 DSH 的正式版上线一并发布正式版! 1. 更新:该插件目前已对齐最新候选版 dsh v0.1.5-rc.1,不适配 Alpha 版。2. 声明:由于该插件编排层依赖官方父子代理交互规则,对于强破坏性变更的版本,不准备适配,因为真的很难修!
★ 28
MIT
TypeScript
Sep 13, 2026
dsh plugin --profile web add dsh-visual-workflowby sandbaseai
Local-first, self-hosted AI agent runtime and MCP bridge with sandboxed sessions, memory, credentials, audit/replay, and a local Console.
★ 647
↓ 8/wk
Apache-2.0
TypeScript
Sep 13, 2026
dsh plugin --profile web add managed-agents