Fast AI Inference - Groq
API capability pack for bounded Groq REST API operations against https://api.groq.com only. Enroll a Groq API key bearer connection named groq; no credential is embedded in this manifest. The pack reads the model catalog, creates non-streaming chat completions, creates bounded Responses API generations, creates text embeddings, and reads batch/file metadata for low-latency LLM workflows, structured output review, semantic-search preparation, model inventory, and operations monitoring. Generation and embedding calls are non-persistent read-tier inference calls that consume Groq quota but do not create durable customer resources through this pack. The pack intentionally excludes streaming, tool calls, tool_choice routing, functions, MCP-connected tools, built-in browser/code/web-search tools, Compound custom tool configuration, arbitrary extra body/query/header passthrough, batch creation or cancelation, file upload/download/delete, model deletion, fine-tuning, multipart audio transcription/translation, binary speech generation, credential management, dynamic outbound callbacks, and any operation outside api.groq.com.
What this pack installs
- Groq chat completion brief Recipe
- Groq Responses API generation brief Recipe
- Groq embedding readiness brief Recipe
- Groq operations digest Recipe
- groq Ingredient
Trust & control
What installing this whole pack would let it do. Recued grants these permissions at install — review them there before approving.