# Open Responses - The NEW Standard API for Open Models Source: https://www.youtube.com/watch?v=b-BzeHF6WLA Recap page: https://rapidrecap.app/video/b-BzeHF6WLA Generated: 2026-01-20T14:14:40.253+00:00 --- ## Quick Overview Open Responses is introduced as a new open-source specification and ecosystem, based on the OpenAI Responses API, designed to provide an interoperable, unified interface for building multi-provider LLM applications, enabling features like consistent streaming, tool use, and reasoning traces across different model providers like OpenAI, Anthropic, Hugging Face, OpenRouter, Vercel, LM Studio, Ollama, and VLLM. **Key Points:** - Open Responses is an open-source specification and ecosystem built around the OpenAI Responses API to create a unified, multi-provider interface for LLMs. - The specification supports key features like consistent streaming events, tool invocation patterns (both externally-hosted and internally-hosted tools), and exposing reasoning traces. - Key community adopters and supporters highlighted include OpenAI, Anthropic, Hugging Face, OpenRouter, Vercel, LM Studio, Ollama, and VLLM. - The framework introduces a state machine for Items with three basic statuses: in_progress, incomplete, and completed, ensuring controlled response flow. - Reasoning items expose the model's internal thought process via content (raw/encrypted) or summary, allowing developers to see model steps or opt for privacy. - The specification standardizes tool calling, allowing models to use tools (like 'get_current_weather') defined in a provider-agnostic manner. - The live demo showcases compatibility with Ollama running locally, successfully executing basic calls, event-based streaming, tool calling, and reasoning traces. ![Screenshot at 00:08: The main title slide clearly presents the project name, "Open Responses," defining it as an open-source specification and ecosystem for building multi-provider, interoperable LLM interfaces based on the OpenAI Responses API.](https://ss.rapidrecap.app/screens/b-BzeHF6WLA/00-00-08.jpg) **Context:** The video introduces Open Responses, a new initiative aiming to standardize the API layer for interacting with various Large Language Models (LLMs). This effort seeks to solve the fragmentation problem where different providers (like OpenAI, Google's Gemini, Anthropic, etc.) use slightly different API schemas for common features such as tool use, streaming, and reasoning. ## Detailed Analysis Open Responses establishes a shared, open specification for LLM interfaces, inheriting structure from the OpenAI Responses API to ensure interoperability across providers. Its design goals include being multi-provider by default with one schema, being friendly to real-world agentic workflows with consistent streaming and tool patterns, and being extensible without fragmentation. The specification defines core concepts like the Agentic Loop, where models iteratively perceive, reason, act through tools, and reflect on outcomes. Items are the fundamental unit of context, tracked through a state machine (in_progress, incomplete, completed). It standardizes tool use, distinguishing between externally-hosted tools (where control yields back to the developer after execution) and internally-hosted tools (where the model executes the tool and returns results without yielding control). Reasoning is exposed via specific item types, providing raw trace, encrypted content, or summaries. The presenter demonstrates successful execution using Ollama locally, showing basic API calls, event-based streaming, tool calling (using a weather function), and reasoning traces, proving the specification's immediate utility across different model types and local/cloud inference solutions. ### Introduction to Open Responses - Open Responses is an open-source specification and ecosystem based on the OpenAI Responses API - Defines a shared schema and tooling layer for unified experience across LLMs - Aims to support multi-provider, agentic workflows. ### Key Principles - Agentic Loop is the core principle, enabling models to perceive input, reason, act through tools, and reflect on outcomes - Items are the fundamental unit of context, carrying atomic model output, tool invocation, or reasoning state - Items follow a state machine: in_progress, incomplete, or completed. ### Specification Details - Reasoning items expose the model's internal thought process using content (raw/encrypted) or summary fields - Tool use is categorized as Externally-hosted tools (control returns to developer) and Internally-hosted tools (model executes tool and returns result directly). ### Tool Calling Standardization - Tool_choice field controls whether the model invokes tools, with options like 'auto', 'required', or 'none' - Demo shows defining a 'get_current_weather' function tool. ### Community Support and Demos - Key supporters include OpenAI, Anthropic, Hugging Face, OpenRouter, Vercel, LM Studio, Ollama, and VLLM - Demos confirm successful execution of basic calls, event-based streaming, tool calling, and reasoning traces using Ollama locally. ![Screenshot at 00:08: The main title slide clearly presents the project name, "Open Responses," defining it as an open-source specification and ecosystem for building multi-provider, interoperable LLM interfaces based on the OpenAI Responses API.](https://ss.rapidrecap.app/screens/b-BzeHF6WLA/00-00-08.jpg) ![Screenshot at 00:37: A comparison slide contrasting 'OLD: SIMPLE CHAT COMPLETIONS' \(limited to conversational text\) with 'NEW: APIS FOR SYSTEMS & AGENTS' \(built for automation, integration, and complex workflows\), highlighting the shift in LLM API focus.](https://ss.rapidrecap.app/screens/b-BzeHF6WLA/00-00-37.jpg) ![Screenshot at 04:05: The community section displaying logos of key supporters and adopters, including OpenRouter, Vercel, Hugging Face, LM Studio, Ollama, OpenAI, and VLLM.](https://ss.rapidrecap.app/screens/b-BzeHF6WLA/00-04-05.jpg) ![Screenshot at 05:27: A section detailing the 'Agentic Loop' principle, showing the continuous cycle of reasoning, tool invocation, and response generation that allows models to gather information before producing a final answer.](https://ss.rapidrecap.app/screens/b-BzeHF6WLA/00-05-27.jpg) ![Screenshot at 11:32: The terminal output from running the demo script against a local Ollama instance, showing successful execution across basic API calls, streaming, and tool calling.](https://ss.rapidrecap.app/screens/b-BzeHF6WLA/00-11-32.jpg)