AG-UI
Open protocol that connects AI agents to frontend applications. Adopted by LangGraph, CrewAI, Mastra, Google ADK, and AWS Bedrock AgentCore. The wiki’s first reference to AG-UI; first appeared in the AIMock launch source as a flagship production user of aimock.
Stub — first appearance in the wiki via the AIMock launch. Will grow when a primary source covers AG-UI directly.
- GitHub: github.com/ag-ui-protocol/ag-ui
Why it’s in the wiki
AG-UI is the frontend-side counterpart to MCP (which connects agents to tools) and A2A (which connects agents to other agents). The three protocols together form the connective layer of the 2026 agentic stack:
| Protocol | Connects | Direction |
|---|---|---|
| MCP | Agents ↔ tools / data sources | Agent reaches out to capabilities |
| A2A | Agents ↔ other agents | Agent collaborates with peers |
| AG-UI | Agents ↔ frontend applications | Agent surfaces to a human user |
The wiki should treat AG-UI as the third leg of this stool. As it gains adoption (already in 5 named frameworks), it becomes the standard way to ship agent UIs — important for any wiki coverage of agent frontend frameworks, CopilotKit specifically, and the broader 2026 “agentic UX” thread.
Production use
AG-UI’s end-to-end test suite uses AIMock with fixture-driven responses across LLM providers. The reference setup at apps/dojo/e2e/fixtures/openai is the canonical “AIMock at scale” example.
See Also
- aimock — used in AG-UI’s E2E tests
- copilotkit — closely affiliated with AG-UI’s development
- mcp — sibling protocol (agent ↔ tools)
- a2a-protocol — sibling protocol (agent ↔ agents)
- langgraph — adoption framework
- Source: AIMock launch (first AG-UI reference)