The Apache-2.0 package is an LLM plugin for models hosted by the Meta AI API. It requires Python 3.10 or later and LLM 0.31.1 or later.
Installation is documented as llm install llm-meta-ai. Users can supply a Meta AI API key with llm keys set meta-ai or the META_AI_TOKEN environment variable. Models use the meta-ai/ prefix; the package documents meta-ai/muse-spark-1.3 and meta-ai/muse-wonder-1.0 examples.
The plugin can list models from the API, caches that list for one hour, and refreshes it with llm meta-ai refresh. It documents reasoning-effort settings, image and PDF attachments, tools, and schemas. Requests count against an output-token rate limit before execution, and a 429 response may mean a listed model is not enabled for a team.
The PyPI page does not provide a changelog or establish what changed specifically from earlier llm-meta-ai releases. Source: https://pypi.org/project/llm-meta-ai/0.2