Skip to main content
Built for real engineering decisions

AI Model & API Guides

Source-linked guidance for AI models, API integration, and developer workflows—turning fast-changing prices, quotas, capabilities, and limits into decisions you can verify.
  • Primary sources first
  • Reproducible steps and failure boundaries
  • Volatile facts carry review dates
  • 482 articles
  • 6 languages
  • 12 topics
  • Updated Oct 7, 2026

An agent execution checkpoint blocks new calls, preserves validated work, and separates uncertain writes for reconciliation.
LatestAPI Guides

How to Stop an AI Agent Tool Loop—and Keep It Stopped

Stop new tool execution, preserve verified results, and reconcile any uncertain write before resuming. A turn limit contains a loop; progress checks and operation records make recovery safer.

12 min
Function-calling migration flow linking application execution and the returned result through the original call_id
API Guides

Migrate Function Calling from Chat Completions to Responses API

Move the whole function-calling loop: validate the model’s arguments, run the authorized handler, return the result with its original call_id, and continue until the user gets a final answer. The JavaScript example preserves typed state and handles failed or unfinished responses.

14 min
Conceptual comparison of uploaded files and local coding tools, organized by execution location and permissions
Developer Tools & Agents

ChatGPT Agent Local Files: Work, Codex, and Claude Code Access

A local path alone grants no access. Upload a selected document, use Work with approved local computer access, or open the checkout in local Codex or Claude Code. Then verify the actual file version and permitted actions.

16 min
Earlier illustration of Imagen 4 and Nano Banana replacement choices
AI Image Generation

Imagen 4 vs Nano Banana: Retirement Dates and API Migration

For a direct Gemini API migration, start with gemini-nano-banana-2.1. Firebase requires a separate support check. Replace the old Imagen request and image parser, then promote the model that meets your output requirements.

12 min
Diagram connecting application JSON and YAML to an image API request, with a separate PDF document-processing step
API Guides

Nano Banana Pro API Guide: JSON, YAML, Image Saving, and PDFs

Call gemini-3-pro-image through Google Interactions, save image bytes from the full response, and compile your own JSON or YAML job settings. PDFs need a document-capable model before Pro receives the verified facts and page images.

12 min
Pro image access differs across Gemini Apps, AI Mode, and native API, followed by separate response-feedback checks.
Troubleshooting

Nano Banana Pro `RESTRICTED by System Instructions`? Check the Gate First

If Pro returns a restriction message or no image, first recover the correct Pro route: More → Redo with Pro in Gemini, the Pro menu in eligible AI Mode accounts, or gemini-3-pro-image in the API. Then check the actual response before calling the failure a policy block.

10 min