openai-responses-api

AI suite powered by state-of-the-art models and providing advanced AI/AGI functions. Includes AI personas, AGI functions, world-class Beam multi-model chats, text-to-image, voice, response streaming, code highlighting and execution, PDF import, presets for developers, much more. Deploy on-prem or in the cloud.

Local AI API gateway for OpenAI / Gemini / Anthropic. Runs on your machine, keeps tokens counted (SQLite), offers priority-based load balancing, optional OpenAI Chat↔Responses format conversion, and one-click setup for Claude Code / Codex.

Self-hosted GitHub Copilot bridge for Claude Code and Codex CLI with OpenAI Responses, Anthropic Messages, streaming resilience, and context management.

Durable local inference for Oh My Pi: NInfer + Qwen3.8 27B on one RTX 3090/4090/5090, with restart-resumable OpenAI Responses state.

Use Claude Code and Codex as the foundation for coding infrastructure, allowing you to decide how to interact with the model with flexible routing and configuration. 中文版： 以 Claude Code 和 Codex 作为编程基础设施的基石，自主选择交互的模型与提供商，并享有灵活的路由与配置。

一个用 TypeScript 构建的多智能体学习与实验场：基于 DeepSeek Responses API，从零实现强类型 Agent Loop、工具调用回填、Skills 按需加载和 SQLite 会话持久化，并以 Gexep、Lexey、Jezeh 等角色“粘合”邮件、E2B Sandbox 等能力，以及对 MCP、A2A、最小权限协作和 GraphRAG 长期记忆的探索与思考。

OCI OpenAI Client for Java simplifies integration between OpenAI’s Java SDK and Oracle Cloud Infrastructure (OCI) GenAI service by providing robust authentication and authorization utilities.

Production‑ready GPT‑6 Astra demo for .NET 10 — Responses API, Microsoft.Extensions.AI, structured JSON, long‑context reasoning, and real-world engineering + BI workflows.

A bare-bones Swift agent runtime with built-in tool execution, skill support, conversation-state handling with a unified multi-provider LLM layer (OpenAI-compatible/Responses, Anthropic, Gemini, and MiniMax, plus codex OAuth)

A small native coding-agent harness with bounded tools, streaming REPL, Agent Skills/Plugins, MCP, and OpenAI-compatible Responses API support.

Protocol-translating LLM proxy for agents

Localhost reverse proxy that bridges OpenAI Codex CLI with Z.ai's GLM Coding Plan API

Local translation proxy: Factory Droid CLI BYOK ↔ any OpenAI-compatible chat-completions endpoint

面向 Docker 部署的 Sub2API 聚合管理控制台，统一管理站点、账户、密钥、注册与 API 服务。

Local LLM serving and measurement tools for Codex CLI on NVIDIA DGX Spark (GB10): a Responses API gateway over vLLM, a vLLM runtime manager for Qwen3.8-Flash-Next, an experimental llama.cpp path for Bonsai 2, and a research CUDA decoder.

Predicts edits from cursor context, diagnostics, and history, offering ghost text and jump hints powered by OpenAI-compatible models.

StarBridge（星桥 / Starbridge / StartBridge）：self-hosted LLM stack for Codex — protocol bridge, KV-aware scheduler, latency tracing（协议桥 + 资源桥 + 观测桥）

Self-hosted multimodal evidence platform with durable ingestion, hybrid/graph retrieval, citations, refusal gates, production validation, Vue, FastAPI, and Docker.

StarBridge（星桥）协议桥 / Codex Adapter：OpenAI Responses <-> Chat bridge with tool preflight and judge-driven retry