AG-UI integrations table entry (draft)

Submitted

Opened as ag-ui-protocol/ag-ui#3280 on 2026-10-31.

How to submit

• Target repository: ag-ui-protocol/ag-ui
• File: README.md, section Clients (the table with the columns Client, Status, AG-UI Resources, Integrations)
• PR title: docs: add reflex-agentkit (Python / Reflex) to the clients table
• When: after reflex-agentkit 0.1.0 is on PyPI, so the link installs a release.
• Follow the repo's contributing guide. If the maintainers prefer the docs site list (supported integrations), reuse the same text there.

The row

Insert it in the Clients table, after the existing community rows:

| [Reflex (Python)](https://github.com/ecrespo/reflex-agentkit) | ✅ Supported | ➡️ [Getting Started](https://github.com/ecrespo/reflex-agentkit#quickstart) 🎮 [Dojo demos](https://github.com/ecrespo/reflex-agentkit/tree/main/examples) | Community |

PR description

This adds reflex-agentkit to the Clients table.

reflex-agentkit is an AG-UI client for Python web apps built with Reflex: a streaming
chat, tool-call cards, human-in-the-loop approvals through standard interrupts
(`RUN_FINISHED` with `outcome.type == "interrupt"`, answered via `RunAgentInput.resume`),
shared state (`STATE_SNAPSHOT` / `STATE_DELTA`) and frontend tools, with no
JavaScript to write.

- It consumes any AG-UI 1.0 server over HTTP + SSE, or runs PydanticAI and LangGraph
  agents in process through their AG-UI integrations.
- It handles all 31 event types of the 1.0 spec.
- The repository ports the five Dojo demos (agentic chat, backend tool rendering, human
  in the loop, shared state, tool-based generative UI) to Python, for both PydanticAI
  and LangGraph.

PyPI: https://pypi.org/project/reflex-agentkit/ · License: MIT