{"id":"googleapis-python-genai","name":"python-genai","homepage":"https://googleapis.github.io/python-genai/","repo_url":"https://github.com/googleapis/python-genai","category":"ai-ml","subcategories":[],"tags":["ai-ml","api","sdk","generative-ai","python","vertex-ai","gemini","client-library"],"what_it_does":"Python SDK for Google generative AI models, providing a client interface (Gemini Developer API and Vertex AI) with Pydantic/dict-based typed request/response structures and helper abstractions (e.g., model content generation, async/sync clients, file upload).","use_cases":["Generate text responses from Gemini models","Generate multimodal outputs (e.g., images) using supported model modalities","Call Gemini/Vertex AI from Python applications with typed request construction","Integrate model invocation into backend services (sync or async)","Upload and summarize files using Gemini Developer API"],"not_for":["Building a non-Python client SDK","Implementing custom model routing without relying on Google APIs","Use cases requiring an embedded local/fully offline LLM (this is a cloud API SDK)"],"best_when":"You want a strongly-typed Python interface to Google GenAI models with either Gemini Developer API or Vertex AI, including async support and convenient content/part modeling.","avoid_when":"You need a REST/GraphQL/grpc service wrapper for your own consumers (this is a client SDK).","alternatives":["Google Cloud Vertex AI SDK for Python","Direct REST calls to Gemini/Vertex AI endpoints","Other GenAI provider SDKs (e.g., OpenAI, Anthropic, AWS Bedrock SDKs)"],"af_score":63.8,"security_score":64.0,"reliability_score":35.0,"package_type":"skill","discovery_source":["openclaw"],"priority":"high","status":"evaluated","version_evaluated":null,"last_evaluated":"2026-03-29T18:03:50.053204+00:00","interface":{"has_rest_api":false,"has_graphql":false,"has_grpc":false,"has_mcp_server":false,"mcp_server_url":null,"has_sdk":true,"sdk_languages":["Python"],"openapi_spec_url":null,"webhooks":false},"auth":{"methods":["API key (GEMINI_API_KEY / GOOGLE_API_KEY)","Vertex AI configuration (project/location with GOOGLE_GENAI_USE_VERTEXAI, GOOGLE_CLOUD_PROJECT, GOOGLE_CLOUD_LOCATION)","Environment-variable based client configuration"],"oauth":false,"scopes":false,"notes":"README describes API-key usage for Gemini Developer API and environment-based Vertex AI selection. It does not specify fine-grained OAuth scopes in the provided content."},"pricing":{"model":null,"free_tier_exists":false,"free_tier_limits":null,"paid_tiers":[],"requires_credit_card":false,"estimated_workload_costs":null,"notes":"Pricing for the underlying Gemini/Vertex AI services is not described in the provided README/manifest content."},"requirements":{"requires_signup":false,"requires_credit_card":false,"domain_verification":false,"data_residency":[],"compliance":[],"min_contract":null},"agent_readiness":{"af_score":63.8,"security_score":64.0,"reliability_score":35.0,"mcp_server_quality":0.0,"documentation_accuracy":80.0,"error_message_quality":0.0,"error_message_notes":null,"auth_complexity":80.0,"rate_limit_clarity":10.0,"tls_enforcement":90.0,"auth_strength":70.0,"scope_granularity":10.0,"dependency_hygiene":70.0,"secret_handling":80.0,"security_notes":"Client communication is designed to use HTTP client libraries (httpx/aiohttp) over network; README emphasizes environment variable configuration for API keys and includes proxy/SSL_CERT_FILE configuration. The provided content does not mention request signing, token scopes, or explicit rate-limit/error-code handling details.","uptime_documented":0.0,"version_stability":60.0,"breaking_changes_history":40.0,"error_recovery":40.0,"idempotency_support":"false","idempotency_notes":null,"pagination_style":"none","retry_guidance_documented":false,"known_agent_gotchas":["The SDK supports multiple API endpoint modes (beta vs stable via http_options api_version); agents should ensure they choose the intended version (v1 vs v1alpha) for reproducibility.","When using custom base_url with base_url_resource_scope, authentication/project/location embedding behavior changes; agents should validate expected request URL construction.","Async usage requires correct client lifecycle management (close/aclose or async context manager) to avoid closed-client errors."]}}