Specialist agent behind a tool
A custom OpenAI agent can run a narrow analysis while General Augment governs when the product assistant may call it.
Category comparison
OpenAI Agents SDK vs an app-agent backend
Use OpenAI Agents SDK when your application owns orchestration, tool execution, approvals, and state. Use General Augment when a customer-facing app needs the managed backend around those turns: stable app users, memory, governed tools, approvals, channels, usage, and traces.
Use OpenAI Agents SDK when
- Your team is writing the agent loop in code.
- The application owns orchestration, tool execution, approvals, and state.
- You want OpenAI-native primitives such as agents, handoffs, guardrails, and tracing.
Use General Augment when
- The agent belongs inside an existing product experience.
- Product teams need memory, tools, approvals, channels, usage, and traces mostly out of the box.
- You want one app-agent backend surface across model providers, tools, channels, and customer-owned connectors.
Decision table
Compare the layer, not just the feature list.
Decision area
OpenAI Agents SDK
General Augment
Abstraction
OpenAI Agents SDK is code-first orchestration for agents, tools, handoffs, guardrails, and tracing.
General Augment is the hosted app-agent backend that product backends call.
State ownership
The app owns state, persistence, approvals, and tool execution around the SDK.
General Augment owns project-scoped memory, approvals, traces, usage evidence, channels, and connector governance.
Best fit
Best for teams building custom OpenAI-native agent systems.
Best for teams adding a governed agent to an existing app without operating the full stack.
Architecture fit
Where the categories fit together.
OpenAI Agents SDK can be the agent implementation layer. General Augment is the app-agent backend layer above model and orchestration choices, with product identity, governance, channels, usage, and traces.
How they can work together
Your app backend
-> General Augment /v1/responses
-> governed tool
-> OpenAI Agents SDK specialist service
-> trace, approval, usage, and final responseSources checked
Migration path
Start where you are. Add the agent backend when the product needs it.
- 01Use OpenAI Agents SDK for custom specialist agents or internal services.
- 02Put customer-facing app-agent turns behind General Augment when identity, memory, tools, approvals, and traces need a product surface.
- 03Expose SDK-built specialist behavior as a governed tool if the product agent should invoke it.
- 04Keep model-provider-specific logic behind server-side boundaries.
Specific examples
Customer-facing assistant
The app backend calls `/v1/responses` with the app user id; General Augment handles memory, approval policy, trace evidence, and channel continuity.
Support review
Operators inspect response ids, traces, tool calls, and memory events without building a custom console around every SDK run.
When not to use General Augment
Use the smaller tool when the smaller tool is enough.
Use OpenAI Agents SDK directly when the agent service is custom infrastructure your team wants to own.
Use OpenAI Agents SDK directly when you are not embedding the agent inside a product with app users and permissions.
Use OpenAI Agents SDK directly when OpenAI-native orchestration is the strategic layer and the rest of the backend is already solved.
FAQ
Does General Augment replace OpenAI Agents SDK?
No. OpenAI Agents SDK is a strong way to build agent behavior in code. General Augment is useful when that behavior needs to become a governed product feature with app users, memory, approvals, channels, usage, and traces.
Can I use both?
Yes. Keep OpenAI Agents SDK for custom specialist orchestration and expose the result through General Augment as a governed product capability.
Which should I start with?
Start with OpenAI Agents SDK if you are building custom agent infrastructure. Start with General Augment if the first milestone is adding an agent backend to an existing app.