tf framework reference — index

The tf framework (import teenyfactories as tf) is the Python library factory agents use to react to data changes, call language models, and persist state. factory.yml defines a factory's states, agents, and volumes.

• tf-common — the core runtime: state-change subscriptions, the dispatch loop, error handling, logging, and the public API surface.
• tf-data — reading, writing, and querying rows in a factory's data store, including vector search.
• tf-llm — calling language models, structured output, agentic tool-calling loops, and embeddings.
• tf-mcp — exposing an agent's capabilities as callable tools.
• tf-environment — secrets, credentials, and configuration available to a running agent.
• tf-volumes — reading and writing files on a factory's attached storage.
• factory-yaml — the factory.yml manifest format: states, agents, and volumes.