Token导航 LogoToken导航TokenDH.com
效率敏感数据clawhub未标认证来源可访问clear审计通过

nano-banana-apinano banana API 文档

Agent Skill

用于辅助 API 设计、接口文档、请求响应结构和服务集成说明。它适合让 Agent 梳理 endpoint、生成 OpenAPI 草稿、检查字段命名、整理错误码或辅助前后端联调。使用时需要确认真实业务语义、鉴权方式、分页和错误处理规则;涉及生成接口文档时,应避免凭空补字段,最好从现有代码、schema 或接口样例中提取事实。

总安装

5,777

周安装

236

GitHub Stars

公开资料未说明

下载量

1,869
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:nano-banana-api(nano banana API 文档)
来源仓库:https://github.com/sdhjkagbjksg/nano-banana-api
安装命令:
openclaw skills install nano-banana-api
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install nano-banana-api

简介

用于辅助 API 设计、接口文档、请求响应结构和服务集成说明。

  • 适合让 Agent 梳理 endpoint、生成 OpenAPI 草稿、检查字段命名、整理错误码或辅助前后端联调。
  • 通过 clawhub 安装并使用 openclaw skills install nano-banana-api 命令部署。
  • 使用时需要确认真实业务语义、鉴权方式、分页和错误处理规则;涉及生成接口文档时,应避免凭空补字段。
  • 可结合来源仓库和原始 README 继续核验具体用法和功能细节。

SKILL.md

name
nano-banana-api
description
Generate images through the Nano Banana REST API and help agents integrate or operate the service safely. Use when a task involves text-to-image, image-to-image, listing models, checking credits, choosing between sync/stream/async generation, polling generation status, or producing reproducible API calls/scripts for Nano Banana.
metadata
openclaw
homepage
https://www.nananobanana.com
primaryEnv
NANO_BANANA_API_KEY
requires
bins
["python3"]

Nano Banana API

Use this skill when an agent needs to call Nano Banana directly instead of only editing docs.

Official website: https://www.nananobanana.com

Quick Start

Set an API key with NANO_BANANA_API_KEY or NB_API_KEY.

Use the bundled CLI for repeatable calls:

python3 scripts/nano_banana_api.py models
python3 scripts/nano_banana_api.py credits
python3 scripts/nano_banana_api.py generate --prompt "A cinematic orange cat on a train" --mode sync
python3 scripts/nano_banana_api.py generate --prompt "Replace the background with a beach" --reference-image-url https://example.com/image.jpg --mode async --wait
python3 scripts/nano_banana_api.py generate --prompt "A cyberpunk skyline at night" --mode stream

Workflow

  1. Call models first when the model name matters.
  2. Omit --model unless the user explicitly asks for a specific model.
  3. Choose request mode by interaction pattern:

- sync: simplest request/response flow. - stream: real-time progress via SSE. - async: immediate submission plus later polling.

  1. Use poll --id <generation_id> --wait when an async task must be followed to completion.
  2. Download images during the same session when the user wants local files. Output URLs expire after 15 days.

Operational Rules

  • Do not print or restate the raw API key.
  • Prefer models over hardcoding catalog values. The live model list changes and is the source of truth.
  • If docs and live API behavior disagree, trust the live endpoints and mention the date of verification.
  • Treat 401, 402, 403, and 503 as actionable operational states:

- 401: invalid or missing key. - 402: insufficient credits. - 403: account does not have API access enabled. - 503: async queue is busy; retry later.

  • For stream, expect newline-delimited SSE data: events.
  • For async, only declare success after processingStatus becomes completed.

Command Guide

List models

python3 scripts/nano_banana_api.py models

Use this before selecting --model. Live verification on 2026-03-25 showed model names that differ from examples in the docs, so avoid assuming nano-banana is the current explicit model ID.

Check credits

python3 scripts/nano_banana_api.py credits

Generate synchronously

python3 scripts/nano_banana_api.py generate \
  --prompt "A watercolor mountain village at sunrise" \
  --mode sync \
  --download-dir ./outputs

Generate asynchronously and wait

python3 scripts/nano_banana_api.py generate \
  --prompt "A product hero shot of a banana-shaped lamp" \
  --mode async \
  --wait \
  --poll-interval 3 \
  --max-polls 60

Follow an existing job

python3 scripts/nano_banana_api.py poll --id clxx123 --wait

Generate from a reference image

python3 scripts/nano_banana_api.py generate \
  --prompt "Keep the subject, turn the room into a brutalist gallery" \
  --reference-image-url https://example.com/source.jpg \
  --mode sync

Fallback Without The Script

If direct shell execution is unavailable, construct HTTP calls from references/api-reference.md. Use:

  • POST /generate for sync, stream, or async submission.
  • GET /generate?id=... for polling.
  • GET /models to resolve current model names.
  • GET /credits to inspect balance.

Resources

适合场景

01

文本生成图片

02

图片风格化

03

产品图和创意图

04

需要 FLUX 模型时

能力概览

能力 1

调用 FLUX 图像模型

能力 2

支持文本生图和图像改写

能力 3

覆盖 LoRA 或风格适配

能力 4

适合创意视觉生成

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

89.63%
按下载量换算1,675

安全审计

VirusTotal

未展示

ClawScan

通过

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills