# Choose an Adapter

Kitaru does not require one agent framework. Pick the boundary that matches the code you already have.

## Decision table

| You have...                 | Use this                | Replay boundary                                         | First page                                                       |
| --------------------------- | ----------------------- | ------------------------------------------------------- | ---------------------------------------------------------------- |
| Plain Python functions      | `@flow` + `@checkpoint` | Your function boundaries                                | [Quickstart](/kitaru/getting-started/quickstart.md)              |
| PydanticAI agent            | `KitaruAgent`           | model/tool/MCP calls by default, or one turn checkpoint | [PydanticAI Adapter](/kitaru/adapters/pydantic-ai.md)            |
| OpenAI Agents SDK agent     | `KitaruRunner`          | per-call checkpoints or one runner-call checkpoint      | [OpenAI Agents Adapter](/kitaru/adapters/openai-agents.md)       |
| Claude Agent SDK invocation | `KitaruClaudeRunner`    | one completed Claude invocation                         | [Claude Agent SDK Adapter](/kitaru/adapters/claude-agent-sdk.md) |
| LangGraph graph             | `KitaruGraphRunner`     | one graph call, or middleware-wrapped model/tool calls  | [LangGraph Adapter](/kitaru/adapters/langgraph.md)               |

## Pick by goal

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><strong>I want the simplest durable boundary</strong></td><td>Wrap phases of work in @checkpoint and keep your existing Python control flow</td><td><a href="/pages/Zt6utJ6bxTsCzHn1d3At">/pages/Zt6utJ6bxTsCzHn1d3At</a></td></tr><tr><td><strong>I want granular agent observability</strong></td><td>Use PydanticAI or OpenAI calls mode so model and tool calls become visible replay units</td><td><a href="/pages/iiOo4wYH4K6fDhJMCaJT">/pages/iiOo4wYH4K6fDhJMCaJT</a></td></tr><tr><td><strong>I want a clean return value from flow.wait()</strong></td><td>Use a coarse runner-call or turn checkpoint around the whole agent turn</td><td><a href="/pages/isuczCq30zR8fJrUKXHc">/pages/isuczCq30zR8fJrUKXHc</a></td></tr><tr><td><strong>I need human approval</strong></td><td>Keep wait/approval bridges at flow scope so the execution can suspend safely</td><td><a href="/pages/BUp6cWRuU8VUfknQKRto">/pages/BUp6cWRuU8VUfknQKRto</a></td></tr></tbody></table>

## Strategy notes

| Strategy                 | Best for                                                            | Trade-off                                                      |
| ------------------------ | ------------------------------------------------------------------- | -------------------------------------------------------------- |
| Coarse checkpoint        | Fast migration, easy `.wait()` return value, low adapter complexity | Replay reruns the whole agent turn if it fails inside the turn |
| Granular checkpoints     | Expensive LLM/tool chains where partial replay matters              | More checkpoint rows and stricter rules around waits/nesting   |
| Explicit raw checkpoints | Maximum control and framework independence                          | You decide every durable boundary yourself                     |

## Human-in-the-loop rule

`kitaru.wait()` belongs at flow scope, not inside a normal checkpoint body. If your harness adapter creates granular tool checkpoints, configure wait-bearing tools so the adapter keeps that wait outside the synthetic checkpoint wrapper.

Start with:

* [Wait, Input, and Resume](/kitaru/guides/wait-and-resume.md)
* [PydanticAI human-in-the-loop tools](/kitaru/adapters/pydantic-ai.md#ask-the-human-from-a-tool-body)
* [OpenAI approval interruptions](/kitaru/adapters/openai-agents.md#approval-interruptions)
* [LangGraph interrupt and resume](/kitaru/adapters/langgraph.md#interrupt-and-resume)

## Next

Run a small adapter example from [Examples](/kitaru/getting-started/examples.md) before porting a production agent. The fastest useful proof is one completed execution where you can see the expected checkpoint shape.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.zenml.io/kitaru/guides/choose-an-adapter.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.