Skip to content

Agent Frameworks

Here you can find the frameworks currently supported in any-agent, along with some basic examples.

Info

If you are missing any agent framework, check the existing issues to see if it has been already requested and comment/upvote on that issue.

If there is no existing issue, don't hesitate to request and/or contribute it.

Examples

Google ADK Repo

from any_agent import AnyAgent, AgentConfig
agent = AnyAgent.create(
    "google",
    AgentConfig(
        model_id="gpt-4o-mini"
    )
)
agent.run("Which Agent Framework is the best??")

LangChain Repo

agent = AnyAgent.create(
    "langchain",
    AgentConfig(
        model_id="gpt-4o-mini"
    )
)
agent.run("Which Agent Framework is the best??")

LLamaIndex Repo

agent = AnyAgent.create(
    "llama_index",
    AgentConfig(
        model_id="gpt-4o-mini"
    )
)
agent.run("Which Agent Framework is the best??")

OpenAI Agents Repo

agent = AnyAgent.create(
    "openai",
    AgentConfig(
        model_id="gpt-4o-mini"
    )
)
agent.run("Which Agent Framework is the best??")

smolagents Repo

agent = AnyAgent.create(
    "smolagents",
    AgentConfig(
        model_id="gpt-4o-mini"
    )
)
agent.run("Which Agent Framework is the best??")

Agno Agents Repo

agent = AnyAgent.create(
    "agno",
    AgentConfig(
        model_id="gpt-4o-mini"
    )
)
agent.run("Which Agent Framework is the best??")

Models

The model used by an agent is defined by 3 arguments model_id, model_type and model_args.

A common usage of model_args is to specify a custom api_base and/or api_key:

from any_agent import AnyAgent, AgentFramework, AgentConfig
agent = AnyAgent.create(
    "smolagents",
    AgentConfig(
        model_id="llama3.2",
        model_args={
            "api_base": "http://localhost:11434/v1"
        }
    )
)
agent.run("Which Agent Framework is the best??")

If you just specify model_id (as in the examples above), the agent will use the default model_type that we have selected for that framework and no model_args.

Tip

For frameworks that have support for LiteLLM (google, langchain, llama_index, smolagents) we use it as default model_type, allowing you to use the same model_id syntax across these frameworks.