AgentWrapper

AgentWrapper/agent-orchestrator

Go Apache-2.0 AI & ML

Agentic orchestrator for parallel coding agents — plans tasks, spawns agents, and autonomously handles CI fixes, merge conflicts, and code reviews.

8.2k stars
1.2k forks
active
GitHub +285 / week

8.2k

Stars

1.2k

Forks

434

Open issues

30

Contributors

AI Analysis

Agent Orchestrator is a meta-harness IDE for supervising multiple AI coding agents (Claude Code, Cursor, Aider, etc.) working in parallel on the same project, each in isolated git worktrees. It automates feedback loops from CI failures, code reviews, and merge conflicts back to the correct agent session. This is a specialized tool for teams or individuals who want to coordinate multiple autonomous coding agents at scale; it is not for general development or single-agent workflows.

AI & ML Application Discovery value: 4/10
Documentation 7/10
Activity 10/10
Community 9/10
Code quality 5/10

Inferred from signals mentioned in the README (tests, CI, type safety) — not a review of the actual code.

Overall score 7/10

AI's overall editorial judgment — not an average of the bars above, can weigh other factors too.

multi-agent-orchestration ai-coding-agents parallel-execution autonomous-workflows agentic-ide
Actively maintained Popular Niche/specialized use case Community favorite Production ready
Deep Analysis · Based on README and public signals
4d ago

Go-based meta-harness for managing parallel AI coding agents across isolated git workspaces

Agent Orchestrator is a local orchestration layer that wraps terminal-based AI coding agents (Claude Code, Codex, Aider, Cursor, etc.) in isolated git worktrees, automatically routing CI failures, review comments, and merge conflicts back to the correct agent session. It targets developers or teams running multiple AI coding agents simultaneously who need a unified supervisor instead of manually tracking parallel terminal sessions and branches. At 8K stars in roughly 5 months, it has attracted meaningful early attention in a rapidly expanding category of agentic development tooling.

Origin

Created February 2026, this project emerged as AI coding agents proliferated and developers began running them in parallel, surfacing a coordination gap that no single CLI or IDE addressed. It is written in Go, distinguishing it from most TypeScript competitors in this space.

Growth

Gained 8,079 stars in approximately 5 months, with 286 stars in the last 7 days suggesting sustained organic momentum rather than a single viral spike. Growth likely correlates with rising adoption of Claude Code, Codex CLI, and similar agent CLIs as professional tools, creating demand for a meta-harness. An active Discord and Twitter presence appear to support community-driven discovery.

In production

Adoption not verified in production environments. The project has community signals (Discord, Twitter, fork count) and mentions support for widely-used agents (Claude Code, Codex, Aider), but no documented case studies, enterprise usage, or quantified install counts are present in the README. Early-adopter usage is plausible given star trajectory, but cannot be confirmed.

Code analysis
Architecture

Appears to follow a daemon-plus-desktop-app architecture: a local daemon watches git worktrees, session state, terminal activity, PRs, CI, and review feedback, while a desktop app and CLI surface state and allow instruction routing. Likely uses git worktrees as the primary isolation primitive. The Go backend likely handles process management, IPC, and the event loop for feedback routing. Architecture description in README is clear, but internal implementation details cannot be verified.

Tests

Not documented in README

Maintenance

Last push was 2026-07-06, same day as analysis — indicating active, current development. The project is under 5 months old with 1,148 forks and an active social presence, suggesting ongoing contributor engagement. No signs of stagnation; cadence appears high for an early-stage project.

Honest verdict

ADOPT IF: you are actively running multiple AI coding agent CLIs in parallel and losing time to manual branch coordination, CI triage, or PR routing — especially if your stack is Go-friendly. AVOID IF: you only run a single agent at a time, your team has no established CI/PR workflow to integrate with, or you need a vetted production-grade tool with documented stability guarantees. MONITOR IF: you are evaluating agentic development workflows and want to track which orchestration layer gains the most ecosystem traction over the next 6–12 months.

Independent dimensions

Mainstream potential

6/10

Technical importance

8/10

Adoption evidence

2/10

Risks
  • Very young project (5 months old) — API surfaces, configuration formats, and core behaviors may change significantly without stability guarantees.
  • No documented production usage or case studies means real-world edge cases (complex monorepos, large teams, flaky CI) are largely untested publicly.
  • Competes in a fast-moving category where well-funded projects (including from AI labs themselves) could absorb this functionality into first-party tooling.
  • Go toolchain requirement may limit adoption in teams whose infrastructure and contributors are TypeScript or Python-centric.
  • Local daemon architecture may introduce reliability concerns (process crashes, state drift) that are not yet addressed by documented recovery or persistence mechanisms.
Prediction

Likely to grow steadily as parallel agent usage becomes standard practice. Risk of consolidation around a smaller number of dominant tools in 12–18 months; AO's trajectory puts it in contention but not yet in a leading position.

0 found this helpful

Newsletter

Get analyses like this every Monday

Free weekly digest of the most interesting open-source discoveries.

Languages

Go
63.4%
TypeScript
28.7%
MDX
5.6%
CSS
1.7%
JavaScript
0.4%
Shell
0.1%
Dockerfile
0%
Nix
0%

Information

Language
Go
License
Apache-2.0
Last updated
7h ago
Created
5mo ago
Analyzed with
anthropic/claude-haiku-4-5

Stars over time

Loading…

Contributors over time

Top 100 contributors only — repos with more will plateau at 100.

Loading…

Similar repos

stablyai

stablyai/orca

Orca is a desktop and mobile application for orchestrating multiple AI coding...

15.3k TypeScript AI & ML
jnMetaCode

jnMetaCode/agency-orchestrator

Agency Orchestrator is a multi-agent AI orchestration platform that automates...

1.7k TypeScript AI & ML
agent-of-empires

agent-of-empires/agent-of-empires

Agent of Empires is a session manager and orchestrator for multiple AI coding...

2.8k Rust Dev Tools
preset-io

preset-io/agor

Agor is a self-hosted, multiplayer web workspace for orchestrating AI coding...

1.3k TypeScript Dev Tools
first-fluke

first-fluke/oh-my-agent

oh-my-agent is a portable multi-agent orchestration harness that enables...

1.2k TypeScript AI & ML
vs. alternatives
stablyai/orca

Orca (12,597 stars, TypeScript) is the current star-count leader in this category. AO differentiates with a Go backend, git worktree isolation as a first-class primitive, and an explicit focus on CI/PR feedback loops rather than just session management.

21st-dev/1code

1code (5,626 stars, TypeScript) appears closer to an AI IDE experience. AO positions itself explicitly as a meta-harness for existing agent CLIs rather than a replacement IDE, which is a narrower but potentially more composable approach.

agent-of-empires/agent-of-empires

Written in Rust (2,761 stars), this project appears to target a similar orchestration use case. Rust vs Go is a meaningful runtime difference for long-running daemons; neither has a clear advantage without more implementation detail.

jnMetaCode/agency-orchestrator

TypeScript-based with 1,661 stars; likely appealing to JS/TS-native teams. AO's Go toolchain may be preferable in environments where Node.js is not the default runtime.

preset-io/agor

agor (1,312 stars, TypeScript) is smaller and less documented publicly. AO has substantially more traction and a more explicit feature surface for CI and PR feedback routing.