Bruto

Product Introduction

1. Definition: Bruto is a local-first, file-based task management and project context tool for software developers and AI coding assistants. It is a web application that runs in the browser and stores all data as a JSON file (.bruto/workspace.json) directly within a project's repository.
2. Core Value Proposition: It exists to solve the context fragmentation problem in AI-assisted development by creating a single, persistent source of truth for tasks, decisions, and rules that is equally accessible to both the human developer and AI agents like Claude Code or Cursor, without requiring cloud accounts or data leaving the local machine.

Main Features

1. Repository-Embedded Task Board: Bruto creates and manages a .bruto/workspace.json file within your project's root directory. This file contains all notes (tasks), their status (todo, in progress, review, done), attached files, screenshots, links, and relationships. This design ensures version control compatibility and direct access for any tool that can read the project folder.
2. AI-Optimized Context Export: The tool provides one-key shortcuts (Q, W, E) to copy a selected note's context to the clipboard. This export includes the note's description, linked files, related notes, and any "standing rules," formatted as clean Markdown with explicit token counts. This feature guarantees the AI receives precise, structured, and complete context for each task.
3. MCP (Model Context Protocol) Server Integration: Bruto can function as an MCP server (bruto-mcp). This allows AI agents with MCP support (e.g., Claude Desktop) to directly interact with the task board using tools. Agents can list notes, read their full context, and post answers back into the note's JSON structure, enabling fully autonomous task execution without manual copy-paste or file editing.
4. Project Structure Visualization & Lenses: Bruto analyzes the project codebase to generate a visual "Structure View." It applies framework-specific lenses for React/Vue/Svelte (showing components and pages), backend APIs like Express/FastAPI (showing resources and endpoints), and databases like PostgreSQL/Prisma (showing tables and schemas), visually mapping notes to the code they affect.
5. Cross-Project Task Dependencies: Notes can block or wait on notes located in a completely different project folder (e.g., a frontend task waiting on a backend API task). Bruto maintains these links, and clicking on a linked note from another project will open that project's board in a new tab, facilitating management of multi-repository workflows.

Problems Solved

1. Pain Point: Context loss and inconsistency when switching between human planning and AI execution. Traditional project management tools (Trello, Jira) or notes apps are not designed for granular, file-attached context that can be seamlessly passed to an LLM.
2. Target Audience: Software developers and engineers using AI coding assistants (Claude Code, Cursor, GitHub Copilot) for daily work; teams managing complex, multi-repository projects; solo developers who need to maintain project context and decisions over time.
3. Use Cases: Providing an AI agent with the exact context for a bug fix, including the error screenshot and relevant file paths. Managing a development sprint where some tasks are handled by Claude Code and others by the developer, with all output tracked in one place. Enforcing coding standards ("standing rules") across all AI-generated code. Coordinating parallel work between separate frontend and backend repositories.

Unique Advantages

1. Differentiation: Unlike cloud-based project management tools, Bruto is local, private, and repository-centric. Unlike generic note-taking apps, it is structurally optimized for developer workflows and AI context transfer. Unlike simple prompt templates, it maintains persistent state, relationships, and a visual map of the project.
2. Key Innovation: The dual-mode operation: a human-friendly GUI for task management and a machine-accessible interface (JSON file + MCP server) for AI agents. This creates a bidirectional collaboration layer where both human and AI work from and contribute to the same persistent data store, bridging the gap between high-level planning and low-level code generation.

Frequently Asked Questions (FAQ)

1. Is Bruto a cloud service or does it store my data online? No, Bruto is a 100% local-first application. It runs in your browser, installs as a PWA, and stores all data in a .bruto/workspace.json file within your project folder. No data is sent to any server except for optional anonymous usage statistics.
2. How does Bruto work with AI agents like Claude Code? You can manually copy task context using the Q/W/E keys to paste into any LLM chat. For deeper integration, Claude Code can directly read and write to the .bruto/workspace.json file. With the MCP server, agents can use dedicated tools to interact with the board without manual file editing.
3. Can I use Bruto with any programming language or framework? Yes, the core task board works with any project. The advanced "Structure View" and auto-detection currently support specific popular frameworks for Web (React, Vue, Svelte, etc.), API (Express, FastAPI, etc.), and Databases (PostgreSQL, Prisma, etc.), but this is an enhancement on top of the universally functional task management.
4. How does Bruto handle merge conflicts if the JSON file is edited by multiple people or an AI? Bruto implements a safe save mechanism. Before writing, it re-reads the file from disk and performs a note-by-note merge of changes. If the file is corrupted or cannot be parsed, it will not overwrite it and will alert the user, maintaining a backup of the last known good state.
5. Is Bruto free and open source? Yes, Bruto is completely free to use and its source code is publicly available under an open-source license (likely MIT) on GitHub, allowing for transparency, community auditing, and self-hosting of the web application.