Genkit Integrations Python

Source of truth

Before proposing, auditing, or editing Genkit code, call search_genkit_docs and then read_genkit_docs for the relevant canonical python/... document. Do not infer package names, signatures, configuration keys, or provider support from memory.

Audit workflow

1. Identify the runtime, Genkit version, plugin versions, environment variables, and execution entry point.
2. Compare each implementation detail against the documents listed below.
3. Report findings by severity with evidence, affected code, fix, and a verification command or test.
4. Do not report style preferences as defects.

Build workflow

1. Choose the smallest documented primitive that satisfies the requirement.
2. Define schemas and error behavior before the implementation.
3. Use the current documented import and initialization pattern.
4. Add an executable verification path using the Dev UI, a test, evaluation, or trace as applicable.

Covered official documents

python/integrations/anthropic.md, python/integrations/deepseek.md, python/integrations/google-cloud.md, python/integrations/google-genai.md, python/integrations/model-providers.md, python/integrations/ollama.md, python/integrations/openai-compatible.md, python/integrations/openai.md, python/integrations/vertex-ai.md, and python/integrations/xai.md.