Busabase logo

Busabase

Organization
busabase

Open-source database & workspace for AI agents — structured data, durable knowledge, reusable skills, runnable apps, and human review on the writes that matter. Local-first and self-hostable.

Publisherbusabase
Repositorybusabase
LanguageTypeScript
Forks
4
Stars
224
Available tools
0
Transport typestdio
Categories
LicenseMIT
Links
  • Connect tools to AI workflows

    Busabase exposes MCP capabilities that can be used by compatible AI clients and agents.

  • 0 available tools

    Browse the callable actions below, including names and descriptions when provided by the server.

  • Ready-to-copy setup

    Use the installation snippets to configure this server in your preferred MCP client.

  • Open source signals

    224 stars and 4 forks from the linked repository.

Busabase is an open-source database and workspace for AI agents — agents and people share the same structured data, docs, skills, and apps, and writes that matter become reviewed, trusted records.

AI agents can write code and generate output, but their useful work usually ends up scattered across chats, files, databases, and SaaS tools. On the next run, the agent has to rediscover the context. When it writes back, humans often cannot see what changed or why.

Busabase gives agents an operational workspace, not another chat window.

  • Database for agents — typed Bases, records, relations, views, forms, and assets.
  • Knowledge base for agents — Docs, Files, Drives, search, history, and provenance.
  • Workspace for agents — Skills, AirApps, Whiteboards, Workflows, Agents, and shared activity.
  • Trust layer for agent work — Change Requests, field-level diffs, comments, approvals, commits, and audit trails.

Agents can read the same workspace you see, use its knowledge and skills, build apps on its data, and write improvements back into it. Every write travels as a Change Request, so it carries a message, a diff, an author, and a full history — and your workspace permissions decide whether it merges on the spot or waits in the Inbox for review.

Free and open source. Local-first. Agent-native. Every change accounted for.

Quick Start

Run it now

bash
npx busabase server

Open http://localhost:15419/dashboard/local. No external database, account, or setup is required. Busabase starts with an embedded PGlite database, local file storage, and demo workspace content.

bash
npm i -g busabase       # install once, then run: busabase server
npx busabase-cli --help # API client for any Busabase server

Docker

bash
docker run --rm -p 15419:15419 -v ~/.busabase/data:/data busabase/busabase

Images are published to Docker Hub (busabase/busabase) and GHCR (ghcr.io/busabase/busabase).

Desktop

Download the native app for macOS, Windows, or Linux at busabase.com/download. Personal Desktop runs locally and works offline; your workspace data does not need to leave your machine.

From source

bash
pnpm install
cp apps/busabase/.env.example apps/busabase/.env
pnpm --filter busabase dev

The local-start check reports missing dependencies or storage configuration before the dashboard opens. The default development setup stores PGlite data under .data/busabase and files under .data/busabase-storage.

Where local data lives

The CLI server and Desktop share one default data root:

text
~/.busabase/data/
├── pgdata/   # embedded PGlite database
└── storage/  # files and attachments

Set BUSABASE_DATA_DIR, PG_DATABASE_URL, or STORAGE_URL to use another location, external Postgres, or S3-compatible storage. Only one process can hold the same PGlite database at a time.

One Workspace, Many Building Blocks

Busabase is not a database with a few AI buttons. Every building block is a first-class node in the same workspace, addressable by humans, agents, MCP, and OpenAPI.

Building blockWhat it gives an agentWhat you get
BaseStructured records, field schemas, relations, filters, and viewsA real operational database instead of unstructured chat memory
DocDurable Markdown knowledge and operating instructionsEditable, versioned knowledge with provenance
File & DriveFiles, attachments, and project treesOne place for the artifacts behind agent work
SkillReusable instructions, references, examples, and scriptsCapabilities that travel with the workspace context
AirAppRunnable apps backed by workspace data and APIsPurpose-built interfaces without creating another silo
Whiteboard & WorkflowVisual context and process definitions agents can readShared plans and processes agents can inspect and improve
Inbox & ActivityProposed changes and workspace eventsHuman control, recovery, and a complete audit trail

Current node types include Folder, Base, Doc, File, Drive, Skill, AirApp, Form, HTML, Whiteboard, and Workflow. See Node Types for the detailed model.

Start from a template

A template installs a whole workspace app at once — its Bases, views, Docs, sample rows, AirApps, and the Skill manual that tells an agent how the app is meant to be used.

bash
busabase-cli install https://github.com/busabase/templates/tree/main/templates/busa-crm

Browse the catalog at busabase.com/templates, or install any Busabase package from a GitHub URL or a local directory. --dry-run prints the plan first, and --require-review leaves the package's records and docs as change requests instead of merging them.

Inside the Workspace

Structured Base for agent dataDurable agent knowledge in a Doc
Database — typed, related, queryable recordsKnowledge base — durable docs with version history
Reusable agent SkillWorkspace-native AirApps
Skills — reusable instructions and supporting filesApps — focused interfaces built on workspace data
Product launch WhiteboardLead intake Workflow
Whiteboards — visual context shared with agentsWorkflows — processes kept beside their data and knowledge
Agent-proposed field diffRecord history and audit trail
Review — see exactly what an agent changedProvenance — see the source, reviewer, commit, and history

On mobile

Review agent Change Requests and open trusted records from the Busabase mobile app.

Connect Your Agent

Busabase has no built-in model. Connect the agent you already use: Claude Code, Codex, Cursor, Gemini CLI, OpenClaw, Hermes, Buda AI, n8n, or your own process.

text
Read and follow the Busabase Agent Skill — it is the single source of truth:
http://localhost:15419/SETUP_SKILL.md

Follow its onboarding to connect to this workspace. Don't choose a merge policy yourself unless I ask for one — submit the change and let Busabase apply my permissions to decide whether it merges now or waits for review. Reply to me in English.

Claude Code guide explains the local skill and Cloud plugin. DeepSeek Harness guide covers the @busabase/dsh-plugin local integration. Bring Your Own Agent covers the agent-neutral flow and permanent skill installation.

Agents can connect in four ways:

ConnectionBest for
Agent SkillCoding agents and local CLIs that can follow workspace instructions
MCPTool-aware agents and IDEs that need typed workspace operations
OpenAPI / CLIApps, scripts, automations, and custom agents
Agents view (ACP)Conversational sessions with inline tool activity and permission requests

Open Agent Skills in the sidebar to get the current setup instructions, MCP endpoint, and OpenAPI specification for your running instance.

Busabase's MCP server is also listed on Glama's MCP connector directory, including its live connector score.

The Trust Loop

Agents need write access to be useful. They also make mistakes. Busabase's answer is not to make every write wait for a human — it is to make every write accountable:

text
Agent reads workspace context
        ↓
Agent writes data, docs, skills, or app changes — always as a Change Request
        ↓
The Change Request carries the diff, the message, the author, and the impact
        ↓
Your permissions decide: it merges on the spot, or it waits in the Inbox
        ↓
Either way the change stays inspectable, attributable, and reversible

Review is a capability, not a toll booth. A credential capped at changeRequest level can only ever propose; a single call can opt in with autoMerge: false; busabase-cli install … --require-review holds a package's content back for approval. Anything else that is allowed to write, writes — and still leaves a diff and a history behind.

This path applies across the workspace. A record update, Doc edit, Skill file, schema change, or AirApp package all carry the same message, diff, merge, and audit history.

What Agents Can Build Here

WorkspaceAgent operationTrusted result
Team memoryCollect notes, sources, decisions, and operating contextA durable knowledge base future agents can reuse
CRM & researchEnrich companies, deduplicate records, monitor marketsVerified business intelligence instead of hidden hallucinations
Content systemDraft posts, docs, pages, metadata, and assetsA headless CMS with an editorial approval trail
Product operationsMaintain projects, tasks, vendors, policies, and configsAn auditable operational database
Dataset pipelineLabel examples, attach evidence, score qualityHuman-reviewed training and evaluation data
Internal toolsBuild AirApps and workflows over workspace dataFocused apps that inherit the same source of truth

See all use cases for complete examples and demo datasets.

A Different Category

Busabase overlaps with databases and knowledge tools, but it is designed around a different primary operator: the agent.

Product categoryPrimary modelMissing when agents do the work
Human databases (Airtable, Baserow, NocoDB)People edit rows directlyAgent context, reusable skills/apps, and a native proposal boundary
Human knowledge tools (Notion, Confluence, Obsidian)People write and organize pagesStructured agent operations across data, files, tools, and review
Databases (Postgres)Applications read and write storageA workspace UI, knowledge model, review loop, and provenance
Agent runtimes and chat toolsAgents execute tasks and produce outputA durable system of record shared across agents and future sessions
BusabaseAgents and humans build one workspace togetherDatabase + knowledge + skills + apps — output that accumulates instead of evaporating

Change Requests are not the category; they are the mechanism that makes an agent workspace dependable.

Personal Desktop and Cloud

Both editions use the same Busabase core and change model.

Personal Desktop / localBusabase Cloud
Open source and freeHosted, multi-user workspace
Local PGlite and file storageManaged Postgres and object storage
No login requiredAuthentication, Spaces, roles, and permissions
Works offlineCollaboration, hosted APIs, and governance
Data stays on your machineAccess from web and mobile

Cloud Connect can link a local workspace to Busabase Cloud through an authenticated tunnel. The local machine keeps the data and can run local agents; Cloud and mobile become controlled windows onto that workspace.

API Surface

Busabase exposes the workspace through MCP, OpenAPI, and busabase-cli. Agents can discover nodes, search content, read file trees, work with records, and create reviewable changes without scraping the UI.

Typical resources include:

  • nodes, folders, and search
  • Bases, fields, views, records, and forms
  • Docs, files, assets, Drives, and Skills
  • AirApps, Whiteboards, Workflows, and HTML nodes
  • Change Requests, operations, reviews, comments, and commits
  • activity, audit events, webhooks, and agent tasks

Open the machine-readable API documentation at:

text
http://localhost:15419/api/v1/doc

Architecture

apps/busabase is the local, single-workspace Next.js shell. The workspace engine lives in packages/busabase-core: nodes, records, file trees, rich node types, review primitives, search, agents, and API contracts. Busabase Cloud runs the same engine with multi-tenant identity, permissions, hosted storage, and collaboration.

The local edition is login-free by construction. It uses a fixed local actor and Space, embedded PGlite by default, and local storage. Cloud supplies real actor and Space context to the shared engine.

Security

The open-source server is designed for a trusted local machine or private network. Do not expose write endpoints directly to the public internet without authentication and a properly configured reverse proxy. Use scoped credentials and Cloud Connect when remote access is required.

Contributing

bash
pnpm install
pnpm --filter busabase dev
pnpm --filter busabase typecheck
pnpm --filter busabase lint:err

Bug reports, feature ideas, docs, and pull requests are welcome in Issues and Discussions.

Community

Star History

License

MIT © Busabase

Use Busabase MCP with multiple AI models

TypingMind connects MCP tools at the workspace level, so once Busabase is connected, you can use it with different AI models in TypingMind instead of setting it up separately for each model. This MCP runs locally through the TypingMind MCP connector on your device.

Setup guide to use the local connector

Use this when the MCP server needs access to local files, apps, or private resources on your computer.

1

Open the MCP settings

In TypingMind, go to Settings, Advanced Settings, then Model Context Protocol and choose Setup Connector.

  1. Open TypingMind in your browser.
  2. Click the Settings icon.
  3. Go to Advanced Settings.
  4. Open the Model Context Protocol section.
  5. Click Setup Connector and choose This Device.
TypingMind MCP connector setup screen with This Device selected
2

Run the connector command

Choose This Device, copy the command from TypingMind, and run it in Terminal. Keep the process running while you use MCP.

  1. Copy the setup command shown by TypingMind.
  2. Open Terminal on macOS or Windows Terminal on Windows.
  3. Paste and run the command.
  4. Approve the package install if Terminal asks you to proceed.
  5. Keep the Terminal window running while using MCP tools.
3

Add Busabase as a server

When the connector status is Ready, click Edit Servers and paste the MCP server configuration.

  1. Wait until the connector status shows Ready.
  2. Click Edit Servers.
  3. Paste the Busabase MCP server configuration.
  4. Save the server list.
  5. Refresh if you want to confirm the connector is still ready.
TypingMind MCP settings showing active server and Edit Servers button
{
  "mcpServers": {
    "busabase": {
      "command": "npx",
      "args": [
        "-y",
        "<mcp-server-package>"
      ]
    }
  }
}
4

Use it across models

Save the server list, open Plugins, enable the Busabase MCP tools, then select any supported AI model in TypingMind and use the tools in chat or assign them to an AI agent.

  1. Open the Plugins page in TypingMind.
  2. Enable the Busabase MCP tools.
  3. Start a chat and choose the AI model you want to use.
  4. Use the MCP tools in chat or assign them to an AI agent.
  5. Switch to another AI model whenever needed without reconnecting MCP.
TypingMind chat using enabled MCP tools with a selected AI model
Can you use Busabase to help me with this task?
Busabase
Sure. I read it.
Here is what I found using Busabase.

Frequently asked questions

What is the Busabase MCP server used for?

Busabase is an MCP server that lets compatible AI clients connect to external tools and context. In TypingMind, you can add this MCP server once and make its tools available in your AI workspace.

Can I use Busabase MCP with multiple AI models in TypingMind?

Yes. TypingMind connects MCP tools at the workspace level, so you can use Busabase with different AI models such as Claude, ChatGPT, Gemini, or other models you have configured in TypingMind without setting up the MCP server separately for each model.

Why use Busabase MCP with TypingMind?

TypingMind is one of the best frontends for LLM chat because it brings multiple AI models, prompts, plugins, AI agents, API keys, and MCP tools into one workspace. With Busabase connected, you can use its MCP tools across your preferred models while keeping your chat workflow organized in TypingMind.

How do I connect Busabase MCP to TypingMind?

Busabase runs through the TypingMind local MCP connector. This is best when the MCP server needs access to local files, desktop apps, command-line tools, or private resources on your computer.

What tools does Busabase MCP provide in TypingMind?

Busabase exposes MCP capabilities that can be enabled from the TypingMind Plugins page and used in chat or assigned to AI agents.

Do I need to share my API keys with TypingMind to use Busabase MCP?

No. TypingMind is local-first and lets you keep your model providers, API keys, prompts, and MCP configuration under your control. If Busabase requires authentication, add the required headers, OAuth settings, or local configuration for that MCP server when you create the connection.

Related MCP Servers

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇