Create an agent with domain-specific knowledge to provide information or take action.
Use Agent Platform
Last verified 22 Jun 2026
Inference provides a single control plane for managing inference workflows. It includes a Model Catalog where you can view available foundation models, including both DigitalOcean-hosted and third-party commercial models, compare model capabilities and pricing, use routing to match inference requests to the best-fit model, and run inference using serverless or dedicated deployments.
Get Started
Use Agent Development Kit to create and manage agents.
Add, edit, or delete API keys to use their models with your agents.
Evaluate Agents
Create test cases and measure how well your agents perform across criteria like factual accuracy, instruction following, and context relevance.
Create evaluation datasets to test agent or model performance, improve accuracy, and measure qualities like factual correctness, safety, and instruction following.
View metrics and runtime logs for your agents to troubleshoot issues.
Test the full agent experience with the directions and features you’ve configured using the Agent Playground.
Use Agents
Use your agent in an application or through a chat bot interface.
Manage Agents
Create workspaces to group agents together and move agents between workspaces as needed.
Integrate multiple generative AI agents.
Enable the foundation model in your agent to access the external data sources using functions.
Rollback to a previous version of an agent to undo changes made to it.
Destroy an agent to permanently and irreversibly destroy the agent and removes all endpoints for the agent.
Manage Knowledge Bases
Create, edit, manage data sources, verify, and permanently destroy knowledge bases.
Attach or detach a knowledge base from your agents.
Test how foundation models answer questions using content retrieved from a knowledge base.
Manage Guardrails
Create, manage, edit, duplicate, or delete guardrails to control how your agents respond to sensitive or inappropriate content.