Local-first MCP server for AI task triage and human approval
taskuary, from Ldbumble, is a local-first Model Context Protocol server that centralizes incoming work requests for automated handling. The app aggregates Email, Microsoft Teams, Slack and report inputs into a unified timeline and routes items to AI agents for triage and execution under user oversight. Key capabilities include multi-channel collection, agentic workflows with MCP-compatible models, and a required approval step. It targets developers, project managers, and power users who need controlled AI-assisted task routing.
Centralizes cross-channel requests into a single AI-ready timeline
Taskuary consolidates Email, Microsoft Teams, Slack and report inputs into one chronological timeline, turning inbound messages into discrete tasks that can be triaged or assigned to agents. The MCP server exposes those items to MCP-compatible models so agentic workflows can act on timeline entries. This reduces the need to jump between communication tools when managing incoming requests during a typical workday.
Automation pairs AI triage with mandatory human approval to limit errors
The app performs automatic categorization and prioritization using AI triage, then allows agents such as Claude Code or Gemini to perform work based on those classifications. Actions produced by agents do not finalize until a user approves them, which keeps the final decision with a human reviewer. For accuracy-sensitive tasks, that approval step separates model output from final execution and limits unchecked automated changes.
Built for developer environments and MCP-compliant hosts
As an MCP server, Taskuary targets Node.js and Python deployment patterns and integrates with MCP-compliant hosts such as Claude Desktop. Setup and agent configuration assume familiarity with hosting, making the app a fit for developers and power users who manage agent credentials and workflows. The project is available as an open-source contribution from ldbumble and can be inspected or adapted by teams that need custom integrations.
Local-first architecture improves data control but external integrations add exposure
Taskuary emphasizes local data handling to enhance privacy, yet it connects to external services like Email, Teams and Slack, which routes messages into the hub for processing. Administrators must configure which data remains local and which is forwarded to MCP agents, so operational decisions affect privacy boundaries. The tool also requires MCP-compatible agents and hosts, which constrains it to that ecosystem rather than every AI platform.
A practical choice for technically capable teams prioritizing controlled AI delegation
Taskuary is a pragmatic option for teams that need controlled delegation of routine tasks to AI agents; its operational demands for hosting and agent configuration narrow the fit to groups with developer resources. The app rewards teams that accept a setup investment with more predictable, reviewable model outputs. Consider this trade-off when planning adoption.
