AI workflow automation tools directory
Compare AI workflow automation tools for agents, integrations, approvals, orchestration, and repeatable business processes.
Filters
Product feed
13 products- ZGI image placeholder
ZGI ingests, cleans, and structures enterprise data and documents, then uses knowledge graphs, multi-recall, and workflow orchestration to build governed AI systems.
Discovered listingFreeDiscovered 
T3 Code — The open-source control plane for coding agents.
Discovered listingFreeDiscovered
Universal provider proxy for OpenAI Codex & Claude Code — use any LLM (Claude, Gemini, Grok, DeepSeek, Ollama…) with Codex CLI, App, SDK, and Claude Code - GitHub - lidge-jun/open.
Discovered listingDiscovered
Hermes Agent is the open-source, self-hosted AI agent by Nous Research.
Discovered listingFreeDiscovered
FreeDiscovered
FreeDiscovered
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs - GitH.
Discovered listingFreeDiscovered
NodeTool is the open-source, agent-first creative workspace.
Discovered listingFreeDiscovered
Multi-agent AI code review that catches real bugs.
Discovered listingFreeDiscovered
You changed the code — did anything break.
Discovered listingFreeDiscovered
Open source, serverless visual workflow automation with AI on Cloudflare.
Discovered listingPaidDiscovered
MemMachine is the open-source memory layer for advanced AI agents, enabling them to learn, remember, and adapt.
Discovered listingFreeDiscovered
Inspect permissions, approvals, and recovery paths
Workflow automation products connect triggers, models, data, and external actions. Compare them on supported integrations, state handling, observability, retry behavior, and the points where a person can approve or stop work.
A successful demo rarely exercises partial failures, duplicate events, expired credentials, or unsafe output. Test those cases before granting production access, use the narrowest permissions available, and confirm that actions are logged and reversible where the underlying system allows it.