#2454·tambo

[FEAT] add Agno (AgnoOS) as a supported agent provider

Author: toliver-hbCreated Feb 22, 2026Updated Feb 24, 2026

Add Agno as a supported agentic framework. Agno exposes agents via the AG-UI protocol through its AgentOS interface, and there is an official TypeScript adapter package @ag-ui/agno (v0.0.3) that extends HttpAgent from @ag-ui/client.

Motivation

We heavily leverage Agno in our own stack and have seen its community growing rapidly. It's a well-designed Python framework for building multi-agent systems with built-in memory, knowledge, and reasoning. It already has first-class AG-UI support via its AGUI interface.

Adding native Tambo support lowers the barrier for that audience and keeps Tambo's framework coverage comprehensive alongside CrewAI, LlamaIndex, Mastra, and PydanticAI.

Proposed Changes

This follows the exact same pattern as the existing CrewAI/LlamaIndex integrations:

  1. packages/core/src/ai-providers.ts: add AGNO = "agno" to AgentProviderType enum
  2. packages/core/src/agent-registry.ts: add registry entry with isSupported: true
  3. packages/backend/src/services/llm/agent-client.ts: add factory case using AgnoAgent from @ag-ui/agno
  4. packages/backend/package.json / apps/api/package.json: add @ag-ui/agno dependency
  5. packages/backend/src/services/llm/agent-client.test.ts: add agent creation test

How Agno's AG-UI adapter works

AgnoAgent is a thin wrapper over HttpAgent:

typescript
// The entire source of @ag-ui/agno
import { HttpAgent } from "@ag-ui/client";

export class AgnoAgent extends HttpAgent {
  public override get maxVersion(): string {
    return "0.0.39";
  }
}

On the Python side, users expose their Agno agent via AgentOS:

python
from agno.agent.agent import Agent
from agno.models.openai import OpenAIChat
from agno.os import AgentOS
from agno.os.interfaces.agui import AGUI

chat_agent = Agent(model=OpenAIChat(id="gpt-4o"))
agent_os = AgentOS(agents=[chat_agent], interfaces=[AGUI(agent=chat_agent)])
app = agent_os.get_app()

References