overstory and agentwise

Both are multi-agent orchestration tools for AI coding agents with a specific focus on Claude Code, making them competitors with B offering additional token optimization, self-improving agents, and automatic verification, while A provides a more general pluggable runtime adapter architecture for broader agent compatibility.

overstory
70
Verified
agentwise
47
Emerging
Maintenance 25/25
Adoption 10/25
Maturity 11/25
Community 24/25
Maintenance 6/25
Adoption 8/25
Maturity 15/25
Community 18/25
Stars: 979
Forks: 183
Downloads:
Commits (30d): 329
Language: TypeScript
License: MIT
Stars: 44
Forks: 12
Downloads:
Commits (30d): 0
Language: TypeScript
License:
No Package No Dependents
No Package No Dependents

About overstory

jayminwest/overstory

Multi-agent orchestration for AI coding agents — pluggable runtime adapters for Claude Code, Pi, and more

Spawns isolated worker agents in git worktrees coordinated via SQLite-backed messaging, merging results through tiered conflict resolution. Supports 11+ pluggable runtimes (Claude Code, Pi, Gemini, Aider, Goose, Amp, etc.) via a unified `AgentRuntime` interface. Includes live TUI dashboards, per-agent inspection, token cost tracking, and explicit warnings about compounding error rates and debugging complexity in multi-agent swarms.

About agentwise

VibeCodingWithPhil/agentwise

Multi-agent orchestration for Claude Code with 15-30% token optimization, self-improving agents, and automatic verification

Implements an 8-agent parallel execution system integrated with Claude Code that coordinates specialized agents for project setup, database configuration, and GitHub automation—eliminating the need for dangerous permission flags through automatic terminal monitoring and sandboxed execution. Extends Claude Code with document upload processing, Figma Dev Mode integration for design-to-code workflows, and website cloning capabilities via MCP server connections. Uses intelligent task distribution with hybrid AI strategy supporting Claude, local models (Ollama, LM Studio), and OpenRouter for optimized cost-performance across different task types.

Scores updated daily from GitHub, PyPI, and npm data. How scores work