Agno Agent API

The /agents folder contains several pre-configured agents:

• Web Search Agent: Performs web searches.
• Agno Assist: Answers questions about Agno. Important: You must load the agno_assist knowledge base via the API before use. The endpoint is: POST /agents/{agent_id}/knowledge/load.
• Finance Agent: Uses the YFinance API for stock prices and financial data.

To get the Agent API running locally, ensure Docker Desktop is installed and running. Follow these steps:

1. Clone the repository:

   git clone https://github.com/agno-agi/agent-api.git
   cd agent-api

2. Configure API keys: The default model is GPT 4.1. Export your OPENAI_API_KEY:

   export OPENAI_API_KEY="YOUR_API_KEY_HERE"

   Note: You can use other providers by updating the agents in the /agents folder.
3. Start the application:

   docker compose up -d

This starts a FastAPI server at http://localhost:8000 and a PostgreSQL database at localhost:5432. You can view the API documentation at http://localhost:8000/docs.

To add or update dependencies in the development environment and ensure they are included in the Docker image:

1. Modify pyproject.toml: Add or update packages in the [dependencies] section.
2. Regenerate requirements.txt: This file is required for the Docker build process.

   ./scripts/generate_requirements.sh

   To upgrade all existing dependencies to their latest compatible versions:

   ./scripts/generate_requirements.sh upgrade

3. Rebuild Docker images:

   docker compose up -d --build

For local development, use uv for package management.

1. Install uv (Unix-like systems):

   curl -LsSf https://astral.sh/uv/install.sh | sh

2. Create environment and install dependencies:

   ./scripts/dev_setup.sh

3. Activate the environment:

   source .venv/bin/activate

   (On Windows, use .venv\Scripts\activate)

The project includes a Dockerfile for production-ready container images.

Build and Push Image

1. Update scripts/build_image.sh with your desired IMAGE_NAME and IMAGE_TAG.
2. Run the build script:

   ./scripts/build_image.sh

Deployment Considerations

• Environment Variables: Securely manage OPENAI_API_KEY and database connection strings in your cloud provider.
• Database: In production, replace the local PostgreSQL container with a managed service (e.g., AWS RDS, Google Cloud SQL, Azure Database for PostgreSQL).
• Port Mapping: The application runs on port 8000 inside the container.
• Supported Platforms: Works with Serverless Container Platforms (Google Cloud Run, AWS App Runner, Azure Container Apps), Orchestration (GKE, ECS, AKS), or PaaS (Railway, Render, Heroku).

Once your local API is running, you can interact with your agents using the Agno Playground:

1. Open Agno Playground.
2. Add http://localhost:8000 as a new endpoint (e.g., name it Agent API).
3. Select the new endpoint to start chatting with your Agents.

The pgvector service provides a PostgreSQL instance with the pgvector extension, which is used for storing vector embeddings. It is configured via the following environment variables:

• POSTGRES_USER: The database user (defaults to ai).
• POSTGRES_PASSWORD: The database password (defaults to ai).
• POSTGRES_DB: The database name (defaults to ai).

Data is persisted in the pgdata volume.

services:
  pgvector:
    image: agnohq/pgvector:16
    environment:
      POSTGRES_USER: ${DB_USER:-ai}
      POSTGRES_PASSWORD: ${DB_PASSWORD:-ai}
      POSTGRES_DB: ${DB_NAME:-ai}

The create_app() function is the factory used to instantiate the FastAPI application. It configures the application title and version from api_settings, sets up the /v1 router, and enables CORS middleware based on the api_settings.cors_origin_list.

If api_settings.docs_enabled is true, the application automatically exposes interactive documentation at /docs (Swagger UI), /redoc (ReDoc), and the OpenAPI schema at /openapi.json.

from api.main import create_app

app = create_app()