AWF

AWF (AI Workflow Framework) is an open-source ecosystem for building, running, and observing AI-powered workflows. It gives developers explicit control over how agents, tools, knowledge, and scheduled tasks work together, from local development to CI pipelines.

At its core, AWF CLI lets you define workflows in version-controlled YAML files. Combine shell commands and AI agents with conditional transitions, parallel execution, loops, retries, and reusable sub-workflows. Connect providers such as Claude, Codex, Gemini, and OpenAI-compatible APIs while keeping your orchestration logic in one place. Execution history, token tracking, and structured logs help you understand what happened and investigate failures.

ZPM adds persistent knowledge and logical reasoning through MCP. Written in Zig and powered by Prolog, it lets agents store facts, apply rules, query relationships, and explain deductions. Named memory segments keep knowledge organized across projects and domains.

Tower gives AI agents a shared, indexed view of your codebase through MCP. Built in Rust, it combines fast file search, code navigation, and version-checked edits, with extensions for semantic analysis, lint fixes, and debugging. A single daemon serves multiple agents within the same workspace, keeping their tools connected to a common index while detecting conflicting edits

ztick handles time-based execution. Applications schedule work when needed, and ztick dispatches it to shell commands, HTTP webhooks, AWF workflows, or messaging systems.

Each component can be used independently or combined into a broader automation stack. Together, they provide a practical foundation for workflows such as code review, documentation generation, and development automation, with readable configuration and traceable execution.