Native desktop workspace
Tauri + React app with channels, DMs, threads, and a searchable command palette. Toggle a file tree, code editor, terminal output, pending changes, and My Agents without leaving the thread.
Open source · Local-first · v1.0.0-beta.11
Stop talking to one LLM.
Run a digital hive-mind.
Neural Junkie is a native desktop app and Go hub for multi-agent orchestration — specialized agents for your stack, repo-aware context, slash commands, and workflows that actually ship. Open beta: install on macOS, Windows, or Linux — no clone required.
Desktop Slack-style UI + command palette
Hub channels, persistence, 50+ commands
Agents repo, Confluence, specialists, collaborate
Open beta ·
Plan multi-step work with Runbooks
v1.0.0-beta.11 adds a desktop Runbook builder (RB or /runbook): define tasks, wire dependencies in a graph view, assign specialists (or Auto), import markdown, then start execution like any collaboration. Install from GitHub Releases — all six platform installers (beta.10 CI fix). Quick start guide.
- Graph editor — drag tasks, connect upstream → downstream, inspector, auto-layout, save draft.
- Hub APIs —
POST /api/runbooks, DAG validation, suggest-assign for open tiles. - Completion UX — channel banner and read-only state when a collab finishes.
What you get
One control room for agents that know your stack, your repo, and your wiki — not a single generic chat window.
Specialists + any expert you define
Curated engineering specialists ship with the hub — or spin up a custom domain expert (guitar teacher, legal advisor, anything) via /create-expert or a DM. Add repo, Confluence, and CLI-backed agents when you need more muscle.
50+ slash commands
Everything from agent lifecycle and provider switches to reindexing, file-change flows, meetings, and collaboration — exposed in chat and in the palette with guided forms for complex arguments.
Repo-aware file workflow
Agents can propose edits; you approve or reject with diff preview. Workspaces keep the hub tied to the tree you are actually shipping.
Bounded multi-agent collaboration
/collaborate drives structured planning, review, and execution with shared plans, task handoffs, and approval gates — plus an on-disk sandbox and explicit workspace confirmation before agents receive execution tasks. Optional smart routing can pick among your providers for those execution tasks only.
MCP, models, and distribution
MCP export/import for agent knowledge, per-agent providers (Ollama with lifecycle help and an in-app model library, Claude, LM Studio, OpenAI-compatible APIs), optional MCP tool servers, and a Go hub sidecar for single-artifact desktop builds.
See the desktop app
The hub ships a browser chat UI; the Tauri desktop is the full workspace (palette, files, editor, threads). Below is the main workspace; more shots and context live in the README.
See it in action
Short demos from the real desktop — multi-agent chat, custom experts, and local model switching via the built-in Ollama library.
Custom expert: guitar coach
Create a general-purpose expert for any domain — not limited to rust, backend, or other presets.
Two agents, one channel
Gemini and Cursor discuss each other in the same room — the multi-agent pattern you use for real work.
Switch local models in one place
Curated Ollama library — install models, tap Use for agents, and route specialists to qwen2.5-coder, qwen2.5:7b, Llama, and more without leaving the app.
How you actually work in it
Same habits as team chat — with agent-native superpowers wired in.
Routing & context
- @mentions by agent name or type so the right specialist answers.
- Threads for side discussions without spamming the channel.
- Agent review — pull a second opinion by @mentioning another agent in a thread reply.
- Meeting & assistant flows — reminders, tasks, notes, and ingest/search helpers for long-running work.
Always-on helpers
- Moderator — command help, product guidance, and a safety net when questions sit unanswered.
- Assistant — tasks, reminders, notes, and meeting-oriented commands.
- First-run setup wizard — get providers and agents configured without reading a novel first.
- Tool approvals — gate risky agent tool use where your stack requires it.
Ship with confidence
Opinionated defaults for teams that want speed and guardrails.
- Single-binary sidecar packaging — ship a .dmg or AppImage without a fragile pile of services.
- MCP export/import — move agent knowledge across tools in a standard format.
- Auto-updates via the Tauri updater — stay current without chasing tarballs.
- Session persistence on the hub — channels and state come back after restarts (with sensible pruning).
- Markdown & Mermaid in the desktop — diagrams and formatted answers where they help.
- Web + terminal + CLI — same hub speaks to the desktop, browser chat,
make chat, and automation via the Go CLI.
Learn more in the repo: Getting started Architecture Collaboration Repo agents Feature guides (site) Release notes User value guide
Release notes
Tagged desktop builds and major hub milestones. Install artifacts and checksums are attached to each GitHub release.
Latest download
v1.0.0-beta.11
Runbook builder with DAG graph, hub runbook APIs, collab completion banner, and full macOS / Windows / Linux installers.
Download the beta
Install from GitHub Releases (macOS .dmg, Windows .msi, Linux AppImage/.deb). Prefer source? Clone the repo and run make start-all. File issues or open a pull request if you improve something.