Google Gemini PaLM API account: setup and migration guide

Setting up a Google Gemini API account gives you access to Google's most capable language models — Gemini 2.0 Flash, Gemini 2.5 Pro, and the broader Gemini family. If you previously used the PaLM API, you need to migrate: PaLM models reached end-of-life, and Google has consolidated everything under the Gemini brand and the Google AI Studio / Vertex AI platforms. CodeWords includes native Gemini access alongside OpenAI and Anthropic — no separate API key setup needed for CodeWords workflows.

PaLM API deprecation timeline

Google's generative AI evolution: PaLM 2 (2023) → Gemini 1.0 (late 2023, PaLM deprecated) → Gemini 1.5, 2.0, 2.5 (2024-2026). Migration path: replace text-bison/chat-bison model names, update the SDK from google-cloud-aiplatform to google-generativeai, and adjust for generate_content() replacing separate generate_text() and chat() methods.

Creating a Google AI Studio API key

Fastest path to Gemini access — no Google Cloud project required: (1) Go to Google AI Studio, (2) Sign in with a Google account, (3) Click "Get API key" → "Create API key", (4) Copy the key. Free tier: Gemini 2.0 Flash at 15 RPM / 1M TPM / 1,500 requests/day; Gemini 2.5 Pro at 2 RPM / 32K TPM / 50 requests/day.

Which model to use

Gemini 2.0 Flash — fastest (~0.5s), cheapest ($0.075/M input tokens), 1M context window. The default choice for most automation use cases: classification, extraction, chatbots, high-volume workflows. Gemini 2.5 Pro — strongest reasoning and coding, $1.25/M input tokens under 200K context. Best for complex analysis, code generation, research synthesis. Gemini 2.5 Flash — balance of speed and capability for mid-complexity tasks.

Vertex AI for production

For higher rate limits, SLAs, data residency compliance, and VPC access: create a GCP project, enable the Vertex AI API, create a service account with the "Vertex AI User" role, download the JSON key, initialize with vertexai.init(project="your-project-id", location="us-central1"). Same Gemini models, enterprise-grade infrastructure. Gemini 2.0 Flash on Vertex: roughly $0.075/M input tokens.

In CodeWords, you access Gemini models alongside OpenAI and Anthropic without managing separate API keys. If Gemini hits rate limits in a CodeWords workflow, the platform routes to OpenAI or Anthropic automatically. Get started with Gemini workflows on CodeWords →