Multica is an open-source AI coding agent harness and orchestration platform developed by multica-ai. It provides a structured runtime architecture designed to orchestrate agent backends, manage skills, and execute automated coding tasks across developer environments. In day-to-day operation, Multica orchestrates underlying AI CLI tools and model engines (including backends like Qwen, Cursor, Copilot, and Pi). It manages agent execution via daemonized processes and structured communication paths, delivering prompts securely over standard streams (such as stdin) to avoid command-line argument mangling and preserve formatting across different operating systems. Multica distinguishes itself through its modular skill system and backend abstraction layer, allowing teams to standardise agent behaviors across different terminal engines and operating systems (with explicit platform optimizations for Windows PowerShell, macOS, and Linux). It includes runtime access controls that allow isolating and hiding private runtimes from non-owners. Multica is distributed as an open-source project hosted on GitHub with an active community and ongoing release cycle.
Multi-Backend Orchestration — Standardizes execution across multiple agent engines and model providers such as Qwen, Copilot, and Cursor. - Robust Process Streaming — Delivers structured prompts over stdin to prevent CLI argument and quote mangling across different platforms. - Extensible Agent Skills — Supports specialized skill modules (such as design guidelines and coding standards) stored directly within the repository structure. - Cross-Platform Invocation Support — Features dedicated execution choosers to ensure reliable invocation across Windows, macOS, and Linux environments. - Runtime Access Control — Manages permissions to isolate private execution runtimes from non-authorized users.
Use cases
Use Case: Cross-Platform AI Agent Orchestration Developers operating in heterogeneous OS environments can run standardized coding agents with uniform prompt delivery and execution behaviors across Windows, macOS, and Linux.
Use Case: Multi-Backend Coding Automation Engineering teams can route complex automated refactoring or feature generation tasks to distinct agent engines (such as Qwen or Copilot) managed under a unified harness.
Use Case: Domain-Specific Agent Skills Teams can package repository-specific workflows and guidelines (such as design system rules) as native agent skills to enforce consistent conventions during automated code generation.
Who it is for
Multica is built for software engineers, DevOps teams, and platform architects looking for an extensible agent harness to manage and orchestrate terminal-based AI coding workflows. It may be overly complex for individual developers looking for a simple, single-binary CLI chat assistant.
Install guide
Installation Clone the repository and build from source: bash git clone https://github.com/multica-ai/multica.git cd multica go build ./...
First Run Ensure your target AI backend (e.g., Qwen CLI or compatible agent tool) is installed and available in your system path. Configure your environment variables as needed for your model backends: bash export MULTICA_RUNTIME=default
Example Session Run an agent task managed by Multica: bash multica run --skill web-design-guidelines "Refactor the navigation header component to match design guidelines"
5dive — 5dive is an open-source agent orchestration harness and management framework developed by 5dive-ai. It provides scaffolding, execution scripts, and infrastructure…
SWE-agent — SWE-agent is an open-source autonomous software engineering agent harness built and maintained by researchers from Princeton University and Stanford University.…
Agentic Engineering Framework — Agentic Engineering Framework is an open-source orchestration harness and governance layer developed by Dimitri Geelen. It is designed to constrain,…
Weaver — Weaver is an open-source agent harness and orchestration framework maintained by sean35mm on GitHub. Built on top of the Bun…
Stratless — Stratless is an open-source agent harness and execution orchestrator developed by the Stratless AI team. It provides a modular infrastructure…
Hivelore — Hivelore is an open-source agent harness and contextual memory orchestrator created by Doucs91. It provides persistent project memory, decision tracking,…