Make
Visual scenario builder with more branching and data-manipulation control than most no-code competitors.
Visit Make →Tools that turn an AI model into something that does a job
A model on its own answers questions. Automation is what makes it complete tasks — pull the record, decide, write it back, tell someone. This directory separates the workflow builders from the agent frameworks from the old-guard RPA vendors, because they solve genuinely different problems and get confused constantly.
24 listings · 5 categories · updated September 2026Advertisement
Browse automation tools
Split by what the tool is really for. Read the category before the logo — a workflow builder and an agent framework are not substitutes for each other.
Microsoft Power Automate
The default choice inside Microsoft-centric organisations, with deep Office, Dynamics and Azure integration.
Visit Microsoft Power Automate →n8n
Source-available workflow automation that can be self-hosted, with native AI nodes and a large integration library. Popular where data cannot leave your infrastructure.
Visit n8n →Pipedream
Developer-first workflows where any step can drop into real code instead of a form.
Visit Pipedream →Tray.ai
General automation platform positioned for technical teams building customer-facing integrations.
Visit Tray.ai →Workato
Enterprise integration and automation platform with governance, environments and approval controls built in.
Visit Workato →Zapier
The broadest integration catalogue in the category, aimed at non-developers connecting SaaS apps with AI steps in between.
Visit Zapier →CrewAI
Framework for multi-agent systems where each agent has a defined role and they collaborate on a task.
Visit CrewAI →Haystack
Production-oriented pipeline framework for search, retrieval and agent applications.
Visit Haystack →LangChain
The most widely used framework for composing model calls, tools and memory into applications, with LangGraph for explicit control flow.
Visit LangChain →LlamaIndex
Focused on connecting models to your own data — ingestion, indexing and retrieval — with agent tooling layered on top.
Visit LlamaIndex →Microsoft AutoGen
Research-led framework for multi-agent conversation patterns, with strong support for human-in-the-loop steps.
Visit Microsoft AutoGen →Semantic Kernel
Microsoft's SDK for embedding model calls and planning into conventional C#, Python or Java applications.
Visit Semantic Kernel →Automation Anywhere
Cloud-native RPA suite with process discovery and AI-assisted bot building.
Visit Automation Anywhere →Blue Prism
Long-established RPA platform with a strong footprint in regulated industries that need auditability.
Visit Blue Prism →Robocorp
Open-source-based, code-first approach to RPA using Python rather than a proprietary designer.
Visit Robocorp →UiPath
The largest RPA vendor, now positioning its robots as the hands for AI-driven decisions in legacy systems.
Visit UiPath →Airflow
The established scheduler for data pipelines, widely repurposed to orchestrate batch AI jobs.
Visit Airflow →Dagster
Asset-oriented orchestration that models what a pipeline produces, not only the steps it runs.
Visit Dagster →Inngest
Event-driven durable functions with retries and step memoisation, aimed at application developers.
Visit Inngest →Prefect
Python-native workflow orchestration with a lighter authoring model than Airflow.
Visit Prefect →Temporal
Durable execution engine that makes long-running, failure-prone workflows resumable. Increasingly used underneath agent systems.
Visit Temporal →Activepieces
Open-source automation platform with AI steps, positioned as a self-hostable alternative to the hosted incumbents.
Visit Activepieces →Model Context Protocol
Open protocol for exposing tools and data to AI agents in a consistent way, so integrations are written once rather than per-client.
Visit Model Context Protocol →No matches. Try a different search.
Advertisement
Questions
Short, direct answers.
What is the difference between a workflow tool and an AI agent?
A workflow tool runs steps you defined, in the order you defined them, and calls a model at specific points. An agent decides its own next step at run time. Workflows are predictable and easy to debug; agents handle situations you did not anticipate, at the cost of being harder to test and reason about. Most production systems that work are mostly workflow with agents used narrowly.
Is RPA obsolete now that language models exist?
No, but its job has narrowed. RPA is still the right answer for driving systems that have no API — clicking through legacy desktop software, for instance. What has changed is that the brittle decision logic bolted onto RPA scripts is often better handled by a model, so the two are increasingly used together rather than one replacing the other.
Do I need a vector database to automate with AI?
Only if the automation needs to search your own unstructured content. If the task is 'take this form and file it correctly', you need integrations and reliable control flow, not retrieval. Vector search is frequently added to projects that never needed it.
What usually breaks first in an AI automation?
The parts that are not AI. Authentication expiring, a downstream API changing its schema, rate limits, and no retry or dead-letter handling account for most production failures. Budget your engineering time for error handling and observability, not for prompt tuning.
Does this directory take payment for listings?
No. Listings are editorial and unpaid. Advertising on the site is labelled and has no bearing on what gets listed or where it appears.