TOOL

302 AI Workbench

Multi-model · BYOK · Developer Integration

AI302.AIMulti-modelBYOKMultimodalAI ChatImage GenerationDeveloper ToolsGPTCodexFree

Open Tool

OVERVIEW

Overview

302 AI Workbench is a local-first, bring-your-own-key multi-model AI tool. Connect your own 302.AI API key to use the models available to your account for multi-turn chat, image understanding, image generation and editing, audio, video, and other multimodal tests. It also includes model browsing, balance and usage tracking, local conversation management, data import/export, and a guided developer integration workflow for GPT / Codex. API keys are not stored by default and are only encrypted and saved in the current browser after explicit user consent.

USAGE

Usage

On first launch, open Connection & Account. If you do not have an API key, jump directly to the 302.AI API Keys page to create one and recharge if needed. Paste the key, verify it, load available models, select a model, and start chatting. To integrate AI into another project, open Developer Integration, choose the capability, project type, and key strategy, download the generated integration package, and send it together with your project files to GPT / Codex.

FEATURES

Features

Unified 302.AI multi-model access with runtime discovery of models available to the current account

Multi-turn AI chat with local conversation save, switch, rename, search, pin, and delete

Multimodal input for images, audio, and video, plus image generation and editing

Markdown rendering with syntax-highlighted code, tables, task lists, and math

Model center with keyword and capability-based filtering

302.AI balance and usage records with per-request details when available

Bring Your Own Key (BYOK): memory-only by default, with optional encrypted browser storage after consent

Chinese and English UI with automatic browser-language detection and manual switching

Developer integration wizard that generates a recommended setup from capability, project type, and key strategy

One-click downloadable integration package designed to be sent directly to GPT / Codex with the project

Conversation history stored in IndexedDB with JSON import, export, and clear-data controls

Pure front-end deployment with no build step or server database required