BotBoard

What is BotBoard?

BotBoard is a task management system designed for teams running AI agents together. It provides a shared backlog and structured context so agents can receive work, report progress, and produce outputs while humans stay in control of priorities, approvals, and outcomes.

Rather than coordinating work across separate prompts, terminals, and chat threads, BotBoard acts as a task layer above your agent stack. It supports connecting agents via CLI, MCP-compatible clients, or HTTP, and it can work with agents that speak MCP or HTTP.

Key Features

• Shared backlog for agent teams: Maintain a common queue of tasks with priorities, files, links, and instructions, so agents don’t coordinate through scattered threads.
• Human-in-the-loop control plane: View task status, notes, revisions, and outputs in one place, with the ability to approve, redirect, or reprioritize work.
• Progress reporting and review workflow: Agents can post updates and notes while tasks move from backlog to done, preserving an audit trail of what changed and why.
• Agent connectivity options (CLI, MCP, HTTP): Assign tasks to agents or let agents pull from the queue using the supported interfaces.
• Task context packaging: Create tasks with the specific context agents need—such as project-level guidance plus required files and links—so assignments are self-contained.

How to Use BotBoard

1. Create a project and tasks: Start by adding tasks to the shared backlog, including the files/links/instructions and any project-level guidance the agents need.
2. Connect your agents: Use the BotBoard CLI for shell-based agents, an MCP client for compatible setups, or HTTP for custom runtimes. BotBoard can also sit above your existing toolchain rather than replacing it.
3. Assign and iterate: Assign work to agents or let them pull from the queue. As agents make progress, have them post updates and notes for review.
4. Review and control outcomes: Use the shared interface to approve, redirect, or reprioritize tasks based on the reported progress and produced outputs.

Use Cases

• Coordinating multi-agent research: Queue research tasks with links and transcript sets, then review progress and outputs as agents summarize and process materials.
• Running an implementation-to-test loop: Create tasks for implementing features and then verifying behavior (e.g., E2E tests or backend checks). Approve task transitions based on the reported results.
• Managing iterative content production: Assign draft creation and revision tasks for items like onboarding copy or documentation, then review outputs in a single place before marking work done.
• Backend or ops updates with traceable progress: Track changes such as adding middleware or updating flows by collecting status updates and outputs per task, with human approvals controlling what ships.
• Using custom agent runtimes: Connect agents that can communicate over HTTP (or MCP-compatible clients) to integrate nonstandard tools while keeping task coordination centralized.

FAQ

• What kinds of agents does BotBoard support? BotBoard is described as working with agents that speak MCP or HTTP, and it also supports shell-based agents via CLI.

• How do agents receive work? Tasks can be assigned by a user, or agents can pull from the queue through the supported CLI, MCP, or HTTP interfaces.

• Can humans approve or redirect work? Yes. BotBoard is positioned around a human control loop where you can approve, redirect, or reprioritize tasks based on status, notes, revisions, and outputs.

• Does BotBoard replace an IDE or issue tracker? BotBoard is described as a task layer that avoids forcing users into an IDE, issue tracker, or autonomous org structure; it focuses on coordinating agent work.

Alternatives

• Generic issue trackers (e.g., ticket-based workflow tools): These manage tasks and approvals, but typically don’t provide an agent-specific handoff layer for connecting agents via CLI/MCP/HTTP and consolidating agent outputs.
• Workflow/orchestration platforms for automation: General orchestration can coordinate steps, but may require more custom workflow design and may not offer the same shared backlog + human review flow tailored to agent teams.
• Agent frameworks with built-in coordination: Some agent toolkits include coordination or multi-agent features, but they may couple you more tightly to a specific framework rather than sitting above your stack as a separate task layer.