uta: A CLI Agent Orchestrator for Parallel AI Task Management
This repository profile is provided by osrepos.com, an open source repository discovery platform.

Summary
uta is a command-line interface (CLI) agent orchestrator designed to decompose complex goals into parallel subtasks. It can dispatch these tasks to various agent CLIs like Claude Code or Gemini CLI, record every step of the process, and synthesize the final results. This tool enhances agent capabilities by enabling parallelism and cross-provider task execution.
Repository Information
Topics
Click on any tag to explore related repositories
Use at your own risk
OSRepos shares public repositories for knowledge and discovery only. Any installation, execution, configuration, or use of code from these repositories is the user's own responsibility. Always review the repository, source code, dependencies, licenses, and security implications before running or installing anything. OSRepos is not responsible for issues, damages, or losses resulting from third-party repositories.
Introduction
uta, short for 'unleash the agents', is a powerful command-line interface (CLI) agent orchestrator. It specializes in breaking down complex goals into independent, parallel subtasks and distributing them to any installed coding-agent CLI, such as Claude Code or Gemini CLI. uta then synthesizes the results and meticulously records every step as a queryable trajectory, providing full transparency and auditability.
Why use and added value
While individual agent CLIs excel at single back-and-forth conversations, uta operates as a crucial layer above, offering significant advantages:
- Parallelism: A built-in planner decomposes goals into independent subtasks, allowing them to run concurrently across multiple providers, significantly speeding up execution.
- Cross-provider Compatibility: Each subtask can target a different CLI, leveraging the strengths of various models, for example, Gemini for large context or Claude for delegation-heavy edits. Custom YAML descriptors allow integration with virtually any other agent.
- Replayable Trajectories: Every prompt, tool call, token count, and final answer is recorded in SQLite. Commands like
uta sessionsanduta trajectory <id>enable inspection, auditing, or resuming any past run. - Modes & Profiles: Mission Profiles allow setting policies per task type (e.g., dev, ops, audit, research), defining allowed tools, human-in-the-loop triggers, budget caps, and retrospective cadences.
- Bidirectional MCP: uta can call Model Context Protocol (MCP) servers as tools and can also expose itself as an MCP server, allowing other MCP-aware clients like Claude Code or Cursor to integrate with uta's planner.
Installation
Installing uta is straightforward, with multiple options available:
- Homebrew (macOS / Linux):
brew install unleashtheagents/tap/uta - curl | sh:
curl -fsSL https://unleashtheagents.ai/install.sh | sh - Go toolchain:
go install github.com/unleashtheagents/uta/cmd/uta@latest
Examples
Get started with uta through a quick 60-second tour of its core functionalities:
uta hello # what is this, what's installed, what to try next
uta doctor # verify environment
uta providers # list detected agent CLIs
uta run -g "summarize this repo in 5 bullets" -y # ad-hoc orchestration
uta sessions # list past runs
uta trajectory <id> # the full event timeline
uta resume <id> -g "now turn each bullet into a tweet"
uta discuss -t "monolith or microservices for this repo?" # two LLMs debate, moderator synthesizes
uta shell # interactive shell: converse with + control agents
uta mission run examples/hello.steer # run a steer program (the agent language, v0)
The uta discuss command enables two agents, ideally different LLMs, to debate a topic over several rounds, with a moderator synthesizing agreements, disagreements, and a final verdict.
uta shell provides a persistent, interactive terminal prompt where you can converse with and control agents. It supports slash commands for model control, tool management, and routing turns to specific agents.
uta mission is the v0 interpreter for steer, a programming language specifically designed for agents. It manages agent calls as effects, enforces linear and mandatory budgets, and journals every run to the session store.
Here's a sample output of a fan-out run:
$ uta run -g "summarize this repo in 5 bullets" -y
[uta] planner=claude/gemini worker=claude max-parallel=3
[uta] plan: 3 subtasks
s1 what does this codebase do?
s2 how is it structured?
s3 what's the install path?
[uta] s1 ? running (claude) ? 1.2s
[uta] s2 ? running (claude) ? 2.1s parallel with s1
[uta] s3 ? running (claude) ? 0.9s parallel with s1/s2
[uta] synthesize ?? claude ? 0.7s
- A CLI for orchestrating multiple agent CLIs in parallel.
- ...
[uta] session 4f6c... status=completed tokens=12,430 approx $0.04
Links
For more detailed information and to contribute, visit the official GitHub repository: uta GitHub Repository
Related repositories
Similar repositories that may be relevant next.

Open Code Review: AI-Powered Code Review Tool from Alibaba
September 18, 2026
Open Code Review is an AI-powered command-line interface tool for code review, battle-tested at Alibaba's scale. It leverages a hybrid architecture, combining deterministic engineering with an LLM agent, to provide secure, fast, and efficient code analysis. This tool generates precise, line-level comments and supports a multi-language ruleset, ensuring high-quality code reviews.

n8n-sandbox-service: Isolated Execution Environments for Your Applications
September 17, 2026
The n8n-sandbox-service provides isolated, on-demand execution environments, leveraging containers or Firecracker microVMs. It offers a REST API to run commands and handle files within these secure sandboxes, ideal for applications requiring ephemeral execution contexts.

Declarative Agents: Profile-Driven LLM Agent Runtime in Go
September 13, 2026
Declarative Agents by Nokia Bell Labs offers a profile-driven runtime and design patterns for building tool-augmented LLM agents. It allows defining agents, their tools, states, and transitions via YAML profiles, eliminating the need for code changes for workflow alterations. This Go-based framework promotes flexible and dependable agent development.

DeclarAgent: Declarative Runbook Executor for Safe AI Agent Workflows
September 12, 2026
DeclarAgent is an innovative declarative runbook executor specifically designed for AI agents. It enables agents to validate, dry-run, and safely execute multi-step YAML workflows. This tool provides a structured, auditable, and secure way for LLM agents to interact with real CLI workflows, enhancing their operational safety and reliability.
Source repository
Open the original repository on GitHub.