AI Agent Platforms / How to choose
How to choose an AI agent platform
There are 49 open-source and source-available platforms in this directory, and they solve different jobs. The fastest way to narrow the field is to answer a few concrete questions in order, rather than comparing every tool at once. This guide walks through those questions and points to the right category for each answer.
1. Start from the job, not the tool
Decide what you actually want to happen. If you want to delegate ongoing work to an agent that uses your tools and returns finished output, look at AI coworkers and teammates. If you are a developer who wants to build custom agents, look at agent builders and frameworks. If you need dependable recurring processes across many services, look at workflow automation platforms. For narrower jobs, there are dedicated coding agents and browser agents.
2. Decide on self-hosting versus cloud
Self-hosting means running the software on your own machine or servers, which keeps your data and tool credentials in your infrastructure. It takes more setup but gives you control and privacy. A managed cloud is faster to start but sends your data to a vendor. Most platforms here can be self-hosted, and many also offer an optional cloud. Each platform page states the exact hosting model.
3. Read the license before you commit
Not everything called open source is licensed the same way. Standard open-source licenses such as MIT and Apache 2.0 allow use, modification, and redistribution with few limits. Source-available and fair-code licenses, such as the Sustainable Use License or a community license, publish the code but restrict some commercial uses. This matters most if you plan to build a product on top of the platform. Every entry in this directory states its exact license, so check it against your intended use.
4. Framework, platform, or ready-to-use product
A framework is a code library you embed and control in code. A platform is a runnable product with its own interface, often with a visual builder and deployment built in. A ready-to-use coworker needs little setup and is managed like a team member. More control usually means more setup. Match the choice to who will build and maintain the agents: an engineering team can use a framework, while a small team without engineers is better served by a product.
5. Check model flexibility
Some platforms lock you to one model provider, while others let you choose the model or bring your own subscription. If you already pay for a model provider, or you expect provider prices and quality to keep changing, prefer a platform that lets you switch. Each platform page lists the documented model support.
6. Consider team needs and operations
If more than one person will use or manage the agents, look for team sharing, access controls, and a way to see what an agent did. For anything running unattended, look for scheduling, logging, and observability. These operational features separate a demo from something you can rely on.
A short decision checklist
- What is the job: delegate work, build agents, or automate a process?
- Must it run on your own infrastructure?
- Does the license fit your intended use, including any commercial plans?
- Who maintains it: engineers who want a framework, or a team that needs a product?
- Can you use the model or subscription you want?
- Do you need team sharing, scheduling, and visibility into agent actions?
Once you have your answers, browse the matching platform pages for documented features, or start from the market map for the whole landscape.