AB
AB
AiBoss

All Categories

project

TinyClaw - TinyAGI's open-source lightweight multi-agent collaboration framework

TinyClaw is a lightweight, open-source multi-agent collaboration framework developed by TinyAGI, designed for resource-constrained environments. TinyClaw supports running multiple specialized AI agents (such as @coder and @writer) simultaneously on devices like Raspberry Pi...

What is TinyClaw?

TinyClaw is a lightweight, open-source multi-agent collaboration framework developed by TinyAGI, designed for resource-constrained environments. TinyClaw supports running multiple professional AI agents (such as @coder and @writer) simultaneously on devices like Raspberry Pi. Agents can collaborate to complete tasks through chained execution and fan-out modes. The system covers three major channels: Discord, WhatsApp, and Telegram, sharing conversation context. A built-in SQLite message queue ensures reliable transmission. Combined with a real-time TUI dashboard and the TinyOffice web backend, it's suitable for building 24/7 edge AI assistant teams.

TinyClaw's main functions

  • Multi-agent systemIt supports running multiple isolated agents simultaneously, each with its own working directory, dialogue history, and configuration file.
  • TeamworkIntelligent agents can sequentially hand over tasks through chained execution or distribute them in parallel to multiple teammates through fan-out mode.
  • Multi-channel accessIt integrates three major platforms: Discord, WhatsApp, and Telegram, allowing all channels to share the same intelligent agent dialogue context.
  • Message QueueIt uses SQLite atomic transaction queues to process messages, and has built-in retry mechanisms and dead letter management to ensure reliable message transmission.
  • Real-time visualizationProvides a TUI terminal dashboard for real-time monitoring of team status, and a TinyOffice Web backend for managing task dashboards.
  • Flexible routingUsers through @agent or @team A prefix can be used to precisely route messages to a specified agent or team for processing.
  • Provider switchingSupports Anthropic Claude and OpenAI Codex, allowing for individual configuration of different AI providers and models for each agent.
  • 7x24 operationIt implements a daemon process mode based on tmux, and works with a heartbeat detection mechanism to ensure long-term stable online operation of the system.
  • Access ControlBuilt-in pairing code verification mechanism; administrators can approve senders to be added to the whitelist to prevent unauthorized users from accessing the proxy.

How to use TinyClaw

  • Installation and DeploymentThe environment deployment can be completed by executing the official one-click installation script. The command is curl followed by the address of the official script, which is then piped to bash for execution.
  • Initialization configurationRun the command `tinyclaw start` to enter the interactive configuration wizard. Follow the prompts to select the required communication channels, configure the AI provider key, name the workspace, and set the default agent.
  • Create an agentUse the tinyclaw agent add command to interactively create specialized role agents. Use tinyclaw agent provider plus agent name, provider and model parameters to specify the AI configuration for each agent individually.
  • Configure TeamThe tinyclaw team add command creates a collaborative team, specifies team members and leaders, and enables task distribution and collaboration between agents.
  • Message routingIn any connected communication channel, use the prefix of @agent_name or @team_name to precisely route messages to the specified handler, such as @coder writing code or @dev fixing bugs.
  • Monitoring and ManagementUse the tinyclaw status command to understand the system's operating status, use the tinyclaw logs all command to troubleshoot problems, or execute tinyclaw team visualize followed by the team name to launch a terminal visualization panel to monitor workflows in real time.
  • Web backendAfter entering the TinyOffice directory, execute the commands to install dependencies and start the development server in sequence. You can then access the visual interface in your browser via a local address to manage task dashboards and agent configurations.
  • Routine maintenanceUse the tinyclaw stop, start, and restart commands to control the service status, upgrade to the latest version using tinyclaw update, or use tinyclaw reset to clear the conversation history but retain the system configuration.

TinyClaw's project address

  • GitHub repositoryhttps://github.com/TinyAGI/tinyclaw

Application Scenarios of TinyClaw

  • Personal intelligent assistantDeploy a personal AI assistant that runs 24/7 on a Raspberry Pi or old computer, receive reminders and manage schedules through multiple channels, and proactively push to-do items using heartbeat detection.
  • Development team collaborationEstablish a code review workflow, where coding agents are responsible for implementing functions, review agents check quality, and documentation agents write explanations simultaneously. The three parties automatically hand over tasks through team mode.
  • Edge device deploymentRun lightweight AI services on resource-constrained embedded devices, with modular design to enable features on demand, maximizing the use of limited computing power to achieve local intelligent processing.
  • Smart Home HubBuild a low-cost smart home control center that can remotely send voice commands via WhatsApp or Telegram to coordinate and control various IoT devices and provide feedback on their execution status.
Back