skills.rest
aiappsgbb
aiappsgbbOFFICIAL
Explore all skills by this author →
GitHub

foundry-doc-vision-speech

Convert images, documents, and audio into structured outputs for Foundry hosted agents.

5|2|Updated Apr 28, 2026
Search Tags:#mcp#vision#foundry#speech-to-text#bicep#agent tools#document intelligence
One-click install
npx skills add https://github.com/aiappsgbb/awesome-gbb --skill foundry-doc-vision-speech
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: foundry-doc-vision-speech
Source: https://github.com/aiappsgbb/awesome-gbb/tree/main/skills/foundry-doc-vision-speech
Command: npx skills add https://github.com/aiappsgbb/awesome-gbb --skill foundry-doc-vision-speech

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill turns unstructured inputs—photos, PDFs, and audio—into structured, tool-ready outputs that a Foundry hosted agent can use to execute the rest of a threadlight process.

Core Features & Use Cases

  • Modality routing & model selection: Chooses the right vision, Document Intelligence, or Speech approach per SPEC guidance (with legacy-model guardrails).
  • Tool contract + Bicep module wiring: Produces the tool contracts and selectors needed for downstream deployment phases and runtime tool invocation.
  • Runtime integration patterns: Supports Foundry Toolbox via MCP and direct SDK fallback for vision, Document Intelligence (prebuilt/custom), and Azure Speech (STT/TTS), including RBAC/keyless constraints.

Example use case: A claim intake SPEC includes a damage photo and a scanned FNOL document plus a voicemail; use this Skill to extract structured fields, transcribe the audio, and feed consistent results into the agent’s action pipeline.

Quick Start

Use the foundry-doc-vision-speech skill when your SPEC requires image analysis, document extraction, or voice intake for a Foundry hosted agent.

Frequently Asked Questions about foundry-doc-vision-speech

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract structured data from images and scanned documents for a Foundry agent?▼

To extract structured data from images and documents, this Skill routes inputs through Azure vision models and Document Intelligence v4. It selects the correct modality stack and generates the required tool contracts for a Foundry hosted agent.

Can I use Azure Speech for voice intake and transcription in an automated process?▼

Yes, Azure Speech supports speech-to-text and text-to-speech intake for threadlight process phases. This Skill wires the necessary runtime integration patterns via Foundry Toolbox MCP or direct SDK fallback.

What is the best way to route unstructured inputs like photos, PDFs, and audio to an agent action pipeline?▼

The best way to route unstructured inputs is to apply modality routing and model selection logic. This Skill evaluates the input type and chooses the appropriate Azure vision, Document Intelligence, or Speech approach.

Does Foundry agent tool integration require keyless RBAC for Document Intelligence and Speech?▼

Yes, Foundry agent tool integration includes keyless RBAC guidance for Document Intelligence and Azure Speech. This Skill provides the necessary constraints and SDK version requirements to maintain secure access.

How do I wire Bicep selectors for downstream deployment when processing damage photos and FNOL documents?▼

To wire Bicep selectors for deployment, this Skill produces the required tool contracts and selectors during the intake phase. It ensures consistent results from damage photos and scanned documents feed into the agent pipeline.

When should I use direct SDK fallback instead of the Foundry Toolbox MCP for vision and speech processing?▼

Use direct SDK fallback when Foundry Toolbox MCP is unavailable or unsupported for your specific runtime. This Skill satisfies both requirements by providing integration patterns and legacy-model guardrails for vision and speech tasks.

Related Skills

Explore More Software Engineering Skills→
truongnattruongnat
1

content-analysis-pro

Extract structured intelligence from documents, images, and video into cited reports.

Community
Advanced
EmeaAppGbbEmeaAppGbb

Hosted Agents in Microsoft Foundry

Deploy and manage containerized AI agents on Microsoft Foundry with azd.

Official
Advanced
aiappsgbbaiappsgbb
5

foundry-prompt-agents

Create and manage Microsoft Foundry prompt agents with model instructions and tool wiring.

Official
Advanced
paidethonpaidethon

azure-ai

Access Azure AI services including search, speech, OpenAI, document intelligence, and content safety.

Community
Advanced