Process automation with AI steps and MCP servers
- License
- MITEnterprise features under a commercial license
- Deployment
- Vendor cloud, Self-hosted, Air-gapped
- Areas
- Build
- Last commit
- 11/2026 · GitHub
60 tools to build, deploy and govern AI agents in one table. Alphabetical, with filters for license and deployment.
At a glance
Mistral Studio, Microsoft Foundry and Gemini platform cover all five. In open source, the same scope usually takes several tools. The “Areas” column in the table shows what each one does.
| Area | What it covers | Examples from the list |
|---|---|---|
| Build | Agents, workflows, connectors to company systems | Dify, Langflow, Sim, Cohere North |
| Evaluate | Experiments, tests, scoring answers with a separate model | Langfuse, Opik, Promptfoo, Braintrust |
| Deploy | Model serving, provider gateways, running agents | vLLM, LiteLLM, Amazon Bedrock AgentCore |
| Govern | Tracing, guardrails, moderation, permissions | NeMo Guardrails, LangSmith, Super Intelligence Control Tower |
| Registry | Versioned catalog of models, prompts and agents | MLflow, Kubeflow, Microsoft Foundry |
Alphabetical list. Narrow it down by license model, deployment, category and area. The “Last commit” column shows whether an open-source project is still maintained. Last update: November 2026.
No tools match these filters.
| Tool | Category | License | Deployment | Areas | Last commit |
|---|---|---|---|---|---|
| Activepieces Activepieces Process automation with AI steps and MCP servers | Workflow and automation | Open core MIT Enterprise features under a commercial license |
|
| GitHub |
| Agent Bricks Databricks Build and govern agents on your Databricks data | Agent platform | Commercial |
|
| — |
| Agenta Agenta Team workspace for building agents and automations | Agent platform | Open core MIT ee/ directories under a separate license |
|
| GitHub |
| Agentforce Salesforce AI agents in the Salesforce ecosystem | Agent platform | Commercial | not stated |
| — |
| Agno Agno Agent framework and runtime with a management UI | Agent platform | Open source Apache 2.0 |
|
| GitHub |
| Amazon Bedrock AgentCore AWS Run production agents on AWS, any framework and model | Agent platform | Commercial |
|
| — |
| AnythingLLM Mintplex Labs Chat and agents over your own documents, runs locally | Team chat | Open source MIT |
|
| GitHub |
| Arize AX Arize AI Agent observability and evaluation | Observability and evals | Commercial |
|
| — |
| Arize Phoenix Arize AI Tracing and evaluation for LLM apps and agents | Observability and evals | Source-available Elastic License 2.0 |
|
| GitHub |
| Braintrust Braintrust Agent observability and evals on production data | Observability and evals | Commercial | not stated |
| — |
| Cohere North Cohere Agent platform that runs where your company data lives | Agent platform | Commercial |
|
| — |
| CrewAI CrewAI Python framework for teams of role-based agents | Framework | Open source MIT On-prem option only in the paid CrewAI AMP |
|
| GitHub |
| Dataiku Dataiku AI and analytics platform with orchestration and governance | MLOps and AI platform | Commercial | not stated |
| — |
| Dify LangGenius Agents, workflows and RAG in one workspace with a visual editor | Agent platform | Source-available Dify Open Source License Apache 2.0 with conditions: no multi-tenant service and no logo removal without permission |
|
| GitHub |
| elsai elsai Agents and governance for regulated-industry workflows | Governance | Commercial |
|
| — |
| Gemini platform Google Cloud · formerly Vertex AI Build, scale and govern agents on Google Cloud | Agent platform | Commercial |
|
| — |
| Google ADK Google Python toolkit to build, test and deploy agents | Framework | Open source Apache 2.0 |
|
| GitHub |
| Guardrails AI Guardrails AI Input and output validation for language models | Guardrails and testing | Open source Apache 2.0 Hosted inference discontinued in 2026 |
|
| GitHub |
| Haystack deepset Framework for RAG pipelines, document processing and agents | Framework | Open source Apache 2.0 |
|
| GitHub |
| Haystack Enterprise Platform deepset Build, run and govern agents, in the cloud or self-hosted | Agent platform | Commercial |
|
| — |
| Helicone Helicone Monitoring of model requests, evals and experiments | Observability and evals | Open source Apache 2.0 |
|
| GitHub |
| IBM watsonx Orchestrate IBM Manage all company agents in one place | Agent platform | Commercial |
|
| — |
| Katonic AI Katonic Sovereign AI platform in your own data center | Agent platform | Commercial |
|
| — |
| Kore.ai Agent Platform Kore.ai Design and manage agents across the enterprise | Agent platform | Commercial | not stated |
| — |
| Kubeflow Kubeflow (CNCF) Set of AI and ML tools running on Kubernetes | MLOps and AI platform | Open source Apache 2.0 |
|
| GitHub |
| Laminar Laminar Tracing for AI agents | Observability and evals | Open source Apache 2.0 |
|
| GitHub |
| Langflow Langflow Visual and code-based agent builder, served as API or MCP server | Agent platform | Open source MIT |
|
| GitHub |
| Langfuse Langfuse Tracing, evals and prompt management for LLM apps | Observability and evals | Open core MIT ee/ directories under a separate license |
|
| GitHub |
| LangGraph LangChain Low-level framework for long-running, stateful agents | Framework | Open source MIT |
|
| GitHub |
| LangSmith LangChain Observability, evals and deployment for agents | Observability and evals | Commercial |
|
| — |
| Letta Letta Agents with persistent memory | Framework | Open source Apache 2.0 Development moved from letta-ai/letta to letta-ai/letta-code |
|
| GitHub |
| LibreChat LibreChat Team chat with agents, MCP and many models | Team chat | Open source MIT |
|
| GitHub |
| LightOn Core LightOn Enterprise AI search over company documents | RAG and search | Commercial | not stated |
| — |
| LiteLLM BerriAI Gateway to 100+ model providers with limits and keys | Gateway and model serving | Open core MIT enterprise/ directory under a commercial license |
|
| GitHub |
| LobeHub LobeHub · formerly LobeChat Chat workspace with agents, knowledge base and plugins | Team chat | Source-available LobeHub Community License |
|
| GitHub |
| Lyzr Lyzr Agent builder deployable in your cloud or on your hardware | Agent platform | Commercial |
|
| — |
| Mastra Mastra TypeScript framework for agent applications | Framework | Open core Apache 2.0 ee/ directories under the Mastra Enterprise License |
|
| GitHub |
| Microsoft Agent Framework Microsoft Framework for building agents in Python and .NET | Framework | Open source MIT |
|
| GitHub |
| Microsoft Foundry Microsoft · formerly Azure AI Foundry Build, ground and govern AI apps and agents on Azure | Agent platform | Commercial |
|
| — |
| Mistral Studio Mistral AI Build, deploy and govern agentic systems | Agent platform | Commercial |
|
| — |
| MLflow Linux Foundation Tracing, evals and a registry for models, agents and prompts | Observability and evals | Open source Apache 2.0 |
|
| GitHub |
| n8n n8n Process automation with AI agent nodes | Workflow and automation | Source-available Sustainable Use License Internal business use only; .ee features need an Enterprise license |
|
| GitHub |
| NeMo Guardrails NVIDIA Programmable safety rails for conversational apps | Guardrails and testing | Open source Apache 2.0 |
|
| GitHub |
| OGX OGX · formerly Llama Stack OpenAI-compatible API server for agents | Agent platform | Open source MIT Project renamed from Llama Stack |
|
| GitHub |
| Open WebUI Open WebUI Chat UI for local and cloud models, works offline | Team chat | Source-available Open WebUI License BSD-3 with a ban on changing branding above 50 users |
|
| GitHub |
| OpenAI Agents SDK OpenAI Python framework for agents with tools, guardrails and tracing | Framework | Open source MIT |
|
| GitHub |
| OpenLIT OpenLIT Cost and behavior monitoring for AI apps | Observability and evals | Open core Apache 2.0 Roles, audit and alerts in OpenLIT Enterprise |
|
| GitHub |
| Opik Comet Tracing, evals and monitoring for LLM apps and agents | Observability and evals | Open source Apache 2.0 |
|
| GitHub |
| Prem Prem AI Sovereign AI deployed in the customer's infrastructure | Agent platform | Commercial |
|
| — |
| Promptfoo Promptfoo (OpenAI) Prompt and agent testing and red teaming | Guardrails and testing | Open source MIT |
|
| GitHub |
| Pydantic AI Pydantic Typed Python framework for agents | Framework | Open source MIT |
|
| GitHub |
| RAGFlow InfiniFlow RAG engine with document layout analysis | RAG and search | Open source Apache 2.0 |
|
| GitHub |
| Red Hat OpenShift AI Red Hat Platform to deploy models and agents across hybrid cloud | MLOps and AI platform | Commercial |
|
| — |
| Sim Sim Shared workspace to build, deploy and monitor agents | Agent platform | Open source Apache 2.0 SSO, SCIM and audit only in Sim Cloud Enterprise |
|
| GitHub |
| StackAI StackAI Build, deploy and govern agents for IT teams | Agent platform | Commercial |
|
| — |
| Super Intelligence Control Tower ServiceNow · formerly AI Control Tower Inventory, security and governance for all AI in the company | Governance | Commercial | not stated |
| — |
| Temporal Temporal Durable execution that retries workflow steps after failures | Workflow and automation | Open source MIT |
|
| GitHub |
| vLLM vLLM High-throughput model serving engine with an OpenAI-compatible API | Gateway and model serving | Open source Apache 2.0 |
|
| GitHub |
| Windmill Windmill Labs Scripts, workflows and agents on a self-hostable orchestrator | Workflow and automation | Copyleft AGPLv3 Enterprise edition and managed service need a commercial license |
|
| GitHub |
| ZenML ZenML ML pipelines and agent evaluation | MLOps and AI platform | Open source Apache 2.0 |
|
| GitHub |
Process automation with AI steps and MCP servers
Build and govern agents on your Databricks data
Team workspace for building agents and automations
AI agents in the Salesforce ecosystem
Agent framework and runtime with a management UI
Run production agents on AWS, any framework and model
Chat and agents over your own documents, runs locally
Agent observability and evaluation
Tracing and evaluation for LLM apps and agents
Agent observability and evals on production data
Agent platform that runs where your company data lives
Python framework for teams of role-based agents
AI and analytics platform with orchestration and governance
Agents, workflows and RAG in one workspace with a visual editor
Agents and governance for regulated-industry workflows
Build, scale and govern agents on Google Cloud
Python toolkit to build, test and deploy agents
Input and output validation for language models
Framework for RAG pipelines, document processing and agents
Build, run and govern agents, in the cloud or self-hosted
Monitoring of model requests, evals and experiments
Manage all company agents in one place
Sovereign AI platform in your own data center
Design and manage agents across the enterprise
Set of AI and ML tools running on Kubernetes
Tracing for AI agents
Visual and code-based agent builder, served as API or MCP server
Tracing, evals and prompt management for LLM apps
Low-level framework for long-running, stateful agents
Observability, evals and deployment for agents
Agents with persistent memory
Team chat with agents, MCP and many models
Enterprise AI search over company documents
Gateway to 100+ model providers with limits and keys
Chat workspace with agents, knowledge base and plugins
Agent builder deployable in your cloud or on your hardware
TypeScript framework for agent applications
Framework for building agents in Python and .NET
Build, ground and govern AI apps and agents on Azure
Build, deploy and govern agentic systems
Tracing, evals and a registry for models, agents and prompts
Process automation with AI agent nodes
Programmable safety rails for conversational apps
OpenAI-compatible API server for agents
Chat UI for local and cloud models, works offline
Python framework for agents with tools, guardrails and tracing
Cost and behavior monitoring for AI apps
Tracing, evals and monitoring for LLM apps and agents
Sovereign AI deployed in the customer's infrastructure
Prompt and agent testing and red teaming
Typed Python framework for agents
RAG engine with document layout analysis
Platform to deploy models and agents across hybrid cloud
Shared workspace to build, deploy and monitor agents
Build, deploy and govern agents for IT teams
Inventory, security and governance for all AI in the company
Durable execution that retries workflow steps after failures
High-throughput model serving engine with an OpenAI-compatible API
Scripts, workflows and agents on a self-hostable orchestrator
ML pipelines and agent evaluation
Licenses
Code on GitHub does not make a project open source. We sort tools into five groups by one rule: whether the license is approved by the Open Source Initiative.
ee/ directory) are under a commercial license..ee require an Enterprise license.Choosing
Checklist
Methodology
A tool makes the list if it helps a company build, test, deploy or govern AI agents and applications built on language models. It needs an official repository or product page that states the license and deployment options.
We do not rank the tools, because we have not tested each one on the same tasks. We do not accept paid listings, and no one pays for a place on the list.
Out of scope: coding agents, personal assistants for a single user (see the list of personal AI agents) and model providers without a platform. Missing a tool, or has something changed? Write to us using the form below.
FAQ
No single project covers that whole scope. The closest are Dify, Sim and Agno. The rest comes from separate tools: Langfuse or Opik for evals and tracing, LiteLLM and vLLM for model deployment, NeMo Guardrails for safety rails.
Dify's code is public, but its license is Apache 2.0 with extra conditions: without the authors' permission you may not run a multi-tenant service or remove the logo. That license is not OSI-approved, so the list puts Dify in the source-available group.
Set the “Deployment” filter to “Air-gapped”. According to the vendors, this mode is available in Red Hat OpenShift AI, Katonic AI, elsai, Open WebUI and Activepieces, among others.
A framework, such as LangGraph, CrewAI or Pydantic AI, is a library for developers: you write the agent in code and deploy it yourself. A platform, such as Dify or Langflow, gives you a ready UI, a runtime and agent management.
We have not tested every tool on the same tasks, so we have no basis to order them from best to worst. We show facts you can check: license, deployment options and project activity.
Once a month. We check whether projects are still maintained and whether licenses changed, and whether commercial products changed their name or deployment options. We also look for new tools. The update date is at the top of the page.
Describe your process and your data requirements. We will help you narrow the list and test the selected tools on your data.
Reply within one business day. No sales call needed.
The form did not load (an ad blocker may be blocking it). Write to us at sales@droptica.com