spinabot/brigade

Brigade — Your personal intelligence, built enterprise-grade

Brigade – a self‑hosted, multi‑agent AI assistant platform

What it is – Brigade is an open‑source ecosystem that lets you run a crew of AI agents on your own machine (or a server/Raspberry Pi). The agents share a long‑term memory store called Tideline, can be organized in a real‑world‑style org chart, and talk to you through many channels (terminal UI, WhatsApp, Telegram, Slack, Discord, iMessage, etc.). All data and API keys stay on your device; no SaaS account or telemetry is required.

Why it matters – It combines several enterprise‑grade ideas—persistent, provenance‑aware memory, fine‑grained access control, and a plug‑in‑style skill system—into a single, install‑once binary. Because the gateway process holds all state and runs continuously, agents can be scheduled, invoked from cron, or triggered by incoming messages, making the system useful for both personal assistants and small‑team automation.

Core features (as described in the README)

Feature What you get
Terminal‑first chat UI A fast, flicker‑free TUI (brigade / brigade tui) that can also act as a thin client for the always‑on gateway.
Tideline long‑term memory File‑backed facts with origin scoping, decay tiers, hybrid keyword + vector recall, and nightly consolidation.
Org‑chart crew Define agents, departments, and reporting lines in cfg.org; policies automatically restrict which agents may talk to each other.
Sub‑agents & fan‑out Spawn isolated child agents (spawn_agent, spawn_agents) with depth and count caps; parent is notified when children finish.
Model‑agnostic Connect to Anthropic, OpenAI, Gemini, OpenRouter, Groq, local Ollama, Claude CLI, etc., and switch models mid‑conversation.
Channel adapters Built‑in adapters for WhatsApp (Baileys), Telegram, Slack (Socket Mode), Discord, iMessage, BlueBubbles; contract ready for more.
1 000+ app connectors Via the Composio SDK you can ask agents to interact with Gmail, GitHub, Notion, Linear, Calendars, etc., using natural‑language commands.
Document & media handling Agents can read PDFs, Office files, images, audio, video and can create/edit Word/Excel/PowerPoint/PDF files using pure‑JS libraries.
HyperFrames video rendering Turn an HTML/GSAP composition into a deterministic MP4 (optional install).
MCP memory server Expose Tideline to any MCP‑compatible client (Claude Desktop, editors, etc.) via brigade mcp.
Self‑hosted storage options Default filesystem under ~/.brigade/ or an optional self‑hosted Convex database with at‑rest encryption.
Cron scheduler Schedule recurring or one‑off jobs that can trigger agent turns, system events, or shell scripts.
Skill system Drop a folder with a SKILL.md into an agent’s workspace; the skill is injected into prompts automatically (56 bundled skills).
Privacy‑first API keys are stored locally with 0600 permissions; no telemetry, no middle‑man, and you can wipe everything with rm -rf ~/.brigade.

Quick‑start flow (from the README)

  1. Install – one‑liner script installs Node (if needed) and the brigade binary, or use npm i -g @spinabot/brigade.
  2. Onboard – brigade onboard walks you through storage mode, provider selection, API‑key entry (or Claude/ChatGPT sign‑in), default model, and web‑search backend.
  3. Run the gateway – brigade gateway run starts the always‑on process; brigade gateway install registers it as a system service.
  4. Chat – brigade tui (or just brigade) opens the terminal chat UI; you can also use brigade connect from other clients or talk via supported messaging platforms.
  5. Extend – add skills, define org hierarchy, schedule cron jobs, or connect Composio connectors.

Who might use it

  • Power users who want a personal AI assistant but refuse cloud‑hosted SaaS and want full control over data.
  • Small teams that need a shared “crew” of agents to automate workflows across many tools (e.g., triaging tickets, drafting reports, posting to Slack).
  • Developers / hobbyists building custom agents, tools, or integrations and needing a plug‑in framework with persistent memory.
  • Enterprises looking for an on‑premise alternative to proprietary AI assistants, because Brigade can run on a Raspberry Pi or a hardened server and integrates with existing identity and storage policies.

Privacy & security highlights

  • Local‑only keys – API secrets never leave the host; they are stored with strict file permissions.
  • No telemetry – The project explicitly states that no usage data is sent to a third‑party service.
  • Access control – Org‑chart policies and origin‑scoped memory prevent agents from leaking information between users or channels.
  • Self‑hosted storage – Choose plain‑filesystem or a self‑managed Convex backend; both keep data under your control.

Where to learn more

Bottom line

Brigade is a genuine, fully‑featured open‑source platform for building, hosting, and interacting with multi‑agent AI crews. It emphasizes data sovereignty, extensibility, and a rich set of integrations while staying usable from a simple terminal UI.

Related

  • Project
    fuyuxiang/echo-agent

    Echo Agent is an open‑source, self‑hosted AI‑agent runtime that provides long‑term memory, automatic skill extraction/evolution, and pluggable tool integration. It can run interactively in a terminal, as a long‑living Gateway service with a web dashboard and API, or be embedded in a desktop application (EchoAgent Desktop). The core runtime manages the agent loop, session context, model routing, permissions, and automation channels, while the memory system recalls past facts and the evolution system turns execution traces into reusable Skills. Install via `pip install "echo-agent[all]"`, configure an LLM API key, and start with `echo-agent run` or `echo-agent gateway`.

  • Project
    dataelement/Clawith

    An open-source multi-agent collaboration platform that treats AI agents as digital employees with persistent identities, long-term memory, and autonomous task management.

  • Project
    peterfei/ai-agent-team

    A virtual AI development team with specialized roles (PM, Backend, Frontend, etc.) and a persistent memory system for managing complex software projects.

  • Project
    topoteretes/cognee

    Cognee is a free, open‑source platform that gives AI agents a persistent, searchable knowledge‑graph memory. It ingests text, code, and other artifacts, stores them locally (or in Postgres), and lets agents retrieve facts or generate answers, optionally using an external LLM. It provides a Python API, CLI, REST, and plugins for popular agent frameworks, and can be run entirely on‑premise with small CPU‑only models.

  • Project
    The-Vibe-Company/companion

    A platform for persistent AI companions with their own isolated Linux runtimes, enabling them to execute code, manage files, and run background automations.