> ## Documentation Index
> Fetch the complete documentation index at: https://phaseo.app/docs/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Phaseo Gateway base URL: https://api.phaseo.app/v1. Send `Authorization: Bearer $PHASEO_API_KEY` with every request.
> Model IDs look like `provider/model`, for example `openai/gpt-6-astra`. Do not guess IDs: list them with `GET /v1/models` or check https://phaseo.app/models. `phaseo/free` routes to an available free model and needs no wallet balance.
> For new text integrations prefer `POST /v1/responses`. `POST /v1/chat/completions` (OpenAI format) and `POST /v1/messages` (Anthropic format) suit existing code.
> The full API contract is https://phaseo.app/docs/openapi/v1/openapi.yaml. Official SDKs: `@phaseo/sdk` (TypeScript) and `phaseo` (Python), plus Go, C#, PHP, Ruby and Rust.
> Search these docs over MCP at https://phaseo.app/docs/mcp, and install Phaseo skills with `npx skills add https://phaseo.app/docs`. If a page is wrong or out of date, report it with the MCP `submit_feedback` tool.
> Manage Phaseo accounts with the `phaseo` CLI (`npm install -g @phaseo/cli`, then `phaseo login` or `phaseo login --device-code`), the `/v1` Management API or the MCP server at `https://mcp.phaseo.app/mcp?access=manage`. Do not send users to the dashboard for account tasks. A person is still needed for sign-up, approving a sign-in, paying the first top-up on the checkout link that `phaseo credits add` returns when no card is saved, confirming each saved-card top-up from the emailed link, and account security (password, MFA, passkeys). Create API keys with `phaseo keys create --name <name> --json`; never guess one. Credit top-ups need `phaseo login --scopes billing:write` first. Task map: https://phaseo.app/docs/v1/developers/manage-from-an-agent.md.

# Extract structured JSON with the Python Agent SDK

> Use the Python Agent SDK for a bounded local-tool loop that returns strict JSON through Phaseo Gateway.

Use this recipe when a Python workflow needs:

* one local lookup tool
* one bounded agent loop
* one strict JSON output shape

## 1. Prepare the SDKs

> [!NOTE]
>
> The Python Agent SDK package is not yet on PyPI. Follow the [source installation instructions](https://phaseo.app/docs/v1/sdk-reference/python/agent-sdk-installation.md).

## 2. Define one local tool

```python theme={null}
from phaseo_agent import AgentTool, define_tool

lookup_policy = define_tool(
    AgentTool(
        id="lookup-policy",
        description="Look up one internal policy by slug.",
        parameters={
            "type": "object",
            "properties": {
                "slug": {"type": "string"},
            },
            "required": ["slug"],
            "additionalProperties": False,
        },
        execute=lambda input, ctx: {
            "slug": input["slug"],
            "summary": "Presets centralize routing, prompts, and caching defaults.",
        },
    )
)
```

## 3. Create the agent

```python theme={null}
from phaseo_agent import create_agent

agent = create_agent(
    {
        "id": "policy-extractor",
        "model": "openai/gpt-5.4-nano",
        "instructions": "Use tools when needed and return concise JSON.",
        "tools": [lookup_policy],
        "parse_output": lambda text: text,
    }
)
```

## 4. Run through the gateway-backed client

```python theme={null}
from phaseo_agent import create_gateway_agent_client

result = agent.run(
    input="Look up presets and return JSON with keys summary and recommendation.",
    client=create_gateway_agent_client(),
)

print(result.output)
```

## 5. When this recipe fits

* one Python worker owns the orchestration
* the tool surface is local and deterministic
* you want strict output without building a larger orchestration layer first

## Related

* [Python Agent SDK](https://phaseo.app/docs/v1/sdk-reference/python/agent-sdk.md)
* [Ship structured JSON from Python](https://phaseo.app/docs/v1/cookbook/sdk-python-presets-and-json.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.