ChatPosit

ChatPosit(
    system_prompt=None,
    base_url='https://gateway.posit.ai',
    model=None,
    credentials=None,
    cache='5m',
)

Chat with a model hosted by Posit AI.

Posit AI provides access to a curated set of models for Posit subscribers. The gateway exposes two API flavors: Claude models are served via the Anthropic Messages API and all other models are served via an OpenAI-compatible API. ChatPosit() automatically picks the appropriate flavor based on the model name.

Prerequisites

Python requirements

Claude models require the anthropic package: pip install "chatlas[posit]". OpenAI-compatible models work with the base chatlas install.

Authentication

By default, ChatPosit() authenticates with an OAuth device flow against login.posit.cloud: the first time you use it, you’ll be prompted to visit a URL and enter a code. The resulting tokens are cached on disk and refreshed automatically, so you should only need to do this once per machine.

Examples

from chatlas import ChatPosit

chat = ChatPosit()
chat.chat("Tell me three jokes about statisticians")

Parameters

Name	Type	Description	Default
system_prompt	Optional[str]	A system prompt to set the behavior of the assistant.	`None`
base_url	str	The base URL of the Posit AI gateway.	`'https://gateway.posit.ai'`
model	Optional[str]	The model to use for the chat. The default, None, will pick a reasonable default, and warn you about it. We strongly recommend explicitly choosing a model for all but the most casual use.	`None`
credentials	Optional[Callable[[], str]]	A zero-argument function that returns a bearer token string. If omitted, `ChatPosit()` manages the OAuth device-flow login and its on-disk token cache automatically.	`None`
cache	Literal['5m', '1h', 'none']	How long to cache inputs? Defaults to “5m” (five minutes). Only applies when a Claude model is selected. See `ChatAnthropic` for details.	`'5m'`

Returns

Name	Type	Description
	Chat	A chat object that retains the state of the conversation. Note that the concrete provider (and hence some response-object typing precision) depends on whether a Claude model or another model was selected.