Inference and Responses - OpenAI
Bounded OpenAI model, Responses, Chat Completions, legacy Completions, embedding, moderation, image, and voice-consent operations. Generated from the current hash-pinned official OpenAI OpenAPI 3.1 document. This leaf carries 17 of the suite's 211 admitted operations. JSON request documents are sent verbatim except where a closed safe subset is required to prevent streaming, credential-bearing nested tools, or variable binary media. Unsafe JSON integers remain exact strings, reads are uncached, sensitive reads are marked sensitive-read, every mutation requires approval, and secret-returning routes stay excluded.
What this pack installs
- openai-inference-responses Ingredient
Trust & control
What installing this whole pack would let it do. Recued grants these permissions at install — review them there before approving.
What it's allowed to do
Destructive
Web APIs
openai-inference-responses
external change
Destructive: audio.delete_voice_consent · Always asksRead: audio.get_voice_consent · Asks approvalRead: audio.list_voice_consents · Asks approvalWrite: audio.update_voice_consent · Always asksRead: chat.create_chat_completion · No approvalDestructive: chat.delete_chat_completion · Always asksRead: chat.get_chat_completion · Asks approvalRead: chat.get_chat_completion_messages · Asks approvalRead: chat.list_chat_completions · Asks approvalWrite: chat.update_chat_completion · Always asksRead: completions.create_completion · No approvalRead: images.create_image · No approvalRead: images.create_image_edit · No approvalDestructive: models.delete_model · Always asksDestructive: responses.cancel_response · Always asksDestructive: responses.delete_response · Always asksRead: untagged.compactconversation · Asks approval