# model-compose - Doramagic AI Context Pack

> Positioning: a pre-install experience and judgment asset. It helps the host AI get off to a good start, but it does not mean the project has already been installed, run, or validated.

## Sufficiency Principle

- **Sufficiency over compression**: The AI Context Pack should be sufficient for the host AI to understand the project's value, capability boundaries, entrypoints, risks, and evidence sources before starting work; it may be layered, but it does not aim for the shortest possible summary.
- **Compression policy**: Compress only noise and duplication, never context that affects judgment or the quality of the work.

## How the Host AI Should Use This

You are reading the AI Context Pack that Doramagic compiled for model-compose. Treat it as pre-work context: help the user understand who it fits, what it can do, how to start, what must be verified after install, and where the risks are. Do not claim that you have already installed, run, or executed the target project.

## Claim Consumption Rules

- **Fact source**: Repo Evidence + Claim/Evidence Graph; the Human Wiki only supplies salience, terminology, and narrative structure.
- **Minimum status for a fact**: `supported`
- `supported`: May be used as a project fact, but the answer must cite the claim_id and evidence path.
- `weak`: Usable only as a low-confidence lead; the user must be asked to keep verifying.
- `inferred`: Usable only for risk notes or open questions; must not be packaged as a project fact.
- `unverified`: Must not be used as fact; state clearly that evidence is insufficient.
- `contradicted`: Must show the conflicting sources and must not force a single version on the user's behalf.

## Who It Fits Best

- **Developers already using host AIs such as Claude/Codex/Cursor/Gemini**: The README or plugin config mentions multiple host AIs. Evidence: `README.md` Claim: `clm_0002` supported 0.86

## What It Can Do

- **Command-Line Startup or Install Flow** (Verify after install): The project documentation contains runnable commands; real use requires running them in a local or host environment. Evidence: `README.md` Claim: `clm_0001` supported 0.86

## How to Start

- `pip install model-compose` Evidence: `README.md` Claim: `clm_0003` supported 0.86
- `git clone https://github.com/hanyeol/model-compose.git` Evidence: `README.md` Claim: `clm_0004` supported 0.86
- `pip install -e .   # or: uv pip install -e .` Evidence: `README.md` Claim: `clm_0005` supported 0.86
- `pip install -e .` Evidence: `README.md` Claim: `clm_0005` supported 0.86, `clm_0006` supported 0.86

## Continue-or-Stop Decision Card

- **Current recommendation**: Needs admin / security approval
- **Why**: Continuing may involve secrets, accounts, external services, or sensitive context; get admin or security approval first.

### 30-Second Read

- **What to do now**: Needs admin / security approval
- **Minimum safe next step**: Run Prompt Preview first; if credentials or an enterprise environment are involved, get approval before trialing
- **Do not trust yet**: Role quality and task fit cannot be trusted directly.
- **Continuing will touch**: Role selection bias, Command execution, Local environment or project files

### What You Can Trust Now

- **Target-audience signal: Developers already using host AIs such as Claude/Codex/Cursor/Gemini** (supported): Backed by a supported claim or project evidence, but that still is not the same as real install results. Evidence: `README.md` Claim: `clm_0002` supported 0.86
- **Capability exists: Command-Line Startup or Install Flow** (supported): You can trust that the project contains signals of this capability; whether it fits your specific task still needs trial or after-install verification. Evidence: `README.md` Claim: `clm_0001` supported 0.86
- **There are Quick Start / install-command signals** (supported): You can trust that the docs mention a startup or install entrypoint; do not run it directly in your primary environment because of that. Evidence: `README.md` Claim: `clm_0003` supported 0.86

### What You Cannot Trust Yet

- **Role quality and task fit cannot be trusted directly.** (unverified): A role library proves there are many roles; it does not prove each one fits your specific task or that a role produces high-quality results.
- **Do not treat role copy as real execution capability.** (unverified): Before install you can only judge whether the role description and task profile match; you cannot prove it can complete the task inside the host AI.
- **Real output quality cannot be trusted before install.** (unverified): Prompt Preview can only show how it guides you; it cannot prove result quality in the real project.
- **Host AI version compatibility cannot be trusted before install.** (unverified): Host loading rules and version differences across Claude, Cursor, Codex, Gemini, and others must be verified in a real environment.
- **That it will not pollute your existing host AI's behavior cannot be trusted directly.** (inferred): Skill, plugin, and AGENTS/CLAUDE/GEMINI instructions may change the host AI's default behavior.
- **Safe rollback cannot be assumed by default.** (unverified): Unless the project clearly provides uninstall and recovery instructions, verify in an isolated environment first.
- **After a real install, is it compatible with the user's current host AI version?** (unverified): Compatibility can only be verified in the actual host environment.
- **Does the project's output quality meet the user's specific task?** (unverified): The pre-install preview can only show flow and boundaries; it cannot replace real evaluation.

### What Continuing Will Touch

- **Role selection bias**: The user's judgment about which expert role should handle the task. Why: Picking the wrong role makes the AI answer from the wrong expert perspective, wasting time or misleading decisions.
- **Command execution**: Package managers, network downloads, the local plugin directory, project config, or the user's home directory. Why: Running the very first command can already change your environment; decide whether it is worth running first. Evidence: `README.md`
- **Local environment or project files**: Install results, plugin caches, project config, or local dependency directories. Why: The write scope and rollback path cannot be proven before install and need isolated verification. Evidence: `README.md`
- **Environment variables / API keys**: Project entry docs explicitly showing API key, token, secret, or account credential configuration. Why: If a real install needs credentials, use test credentials first and go through a permission/compliance review. Evidence: `README.ko.md`, `README.md`, `README.zh-cn.md`, `docs/user-guide/04-component-configuration.md` et al.
- **Host AI context**: The AI Context Pack, Prompt Preview, Skill routing, risk rules, and project facts. Why: Importing context affects the host AI's later judgment, so avoid packaging unverified items as facts.

### Minimum Safe Next Steps

- **Run Prompt Preview first**: Use an interactive trial to verify the task profile and role match first; do not import the whole role library up front. (applies when: Applies to any project, especially when output quality is unknown.)
- **Trial-install only in an isolated directory or a test account**: Avoid letting install commands pollute your primary host AI, real projects, or home directory. (applies when: When there are signals of command execution, plugin config, or local writes.)
- **Do not use real production credentials**: Once an environment variable / API key enters the host or toolchain, it can create account and compliance risk. (applies when: When environment signals like API, TOKEN, KEY, or SECRET appear.)
- **After install, verify just one minimal task**: Verify loading, compatibility, output quality, and rollback first, then decide whether to use it deeply. (applies when: When moving from a trial into a real workflow.)

### Exit Plan

- **Preserve the pre-install state**: Record the original host config and project state so you can later judge whether it is recoverable.
- **Keep a record of the original role selection**: If output goes off-topic, you can return to the task-profiling stage and reselect a role instead of pushing on with the wrong one.
- **Record the install commands and written paths**: Without clear uninstall instructions, you at least need to know which directories or configs to clean up manually.
- **Be ready to revoke test API keys or tokens**: If test credentials leak or are misused, you can cut losses quickly.
- **If there is no rollback path, do not enter your primary environment**: No rollback is a blocker before continuing; do not proceed on trust or luck.

## What Can Only Be Previewed

- Explain who the project fits and what it can do
- Demonstrate a typical conversation flow based on project docs
- Help the user decide whether it is worth installing or researching further

## What Must Be Verified After Install

- Actually installing the Skill, plugin, or CLI
- Running scripts, modifying local files, or accessing external services
- Verifying real output quality, performance, and compatibility

## Boundary & Risk Decision Card

- **Mistaking the pre-install preview for a real run**: The user may overestimate how much configuration, permission, and compatibility verification the project has already done. Mitigation: Clearly separate prompt_preview_can_do from runtime_required. Claim: `clm_0007` inferred 0.45
- **Command execution will modify the local environment**: Install commands may write to the user's home directory, the host plugin directory, or project configuration. Mitigation: Run in an isolated environment or a test account first. Evidence: `README.md` Claim: `clm_0008` supported 0.86
- **To confirm**: After a real install, is it compatible with the user's current host AI version?. Why: Compatibility can only be verified in the actual host environment.
- **To confirm**: Does the project's output quality meet the user's specific task?. Why: The pre-install preview can only show flow and boundaries; it cannot replace real evaluation.
- **To confirm**: Do the install commands require network access, permissions, or global writes?. Why: This affects install risk in both enterprise and personal environments.

## Pre-Work Working Context

### Loading Order

- First read how_to_use.host_ai_instruction to establish the boundaries of this pre-install judgment asset.
- Read claim_graph_summary to confirm facts come from the Claim/Evidence Graph, not the Human Wiki narrative.
- Then read intended_users, capabilities, and quick_start_candidates to judge whether the user is a match.
- When you need to carry out a concrete task, check role_skill_index first, then evidence_index.
- For real install, file modification, network access, performance, or compatibility questions, turn to risk_card and boundaries.runtime_required.

### Task Routes

- **Command-Line Startup or Install Flow**: State that this is an after-install capability first, then give a pre-install checklist. Boundary: Must be verified after a real install or run. Evidence: `README.md` Claim: `clm_0001` supported 0.86

### Context Scale

- Total files: 1381
- Important-file coverage: 40/1381
- Evidence index entries: 80
- Role / Skill entries: 80

### Handling Insufficient Evidence

- **missing_evidence**: State that evidence is insufficient and ask the user for the target file, a README section, or after-install verification records; do not fill in facts.
- **out_of_scope_request**: State that the task is beyond the current AI Context Pack's evidence scope and suggest the user check the Human Manual or verify after a real install.
- **runtime_request**: Provide a pre-install checklist and command sources, but do not run commands for the user or claim they have been run.
- **source_conflict**: Show the conflicting sources side by side, mark them as unverified, and do not force a single version.

## Prompt Recipes

### Fit assessment

- Goal: Judge whether this project fits the user's current task.
- Expected output: A fit conclusion, key reasons, evidence citations, what can be previewed before install, what must be verified after install, and a next-step recommendation.

```text
Based on the AI Context Pack for model-compose, ask me 3 necessary questions first, then judge whether it fits my task. The answer must cover: who it fits, what it can do, what it cannot do, whether it is worth installing, and where the evidence comes from. Every project fact must cite evidence_refs, source_paths, or a claim_id.
```

### Pre-install experience

- Goal: Let the user feel the core workflow before installing, while avoiding packaging the preview as real capability or a marketing promise.
- Expected output: An experience script with boundary labels, an after-install verification checklist, and a cautious recommendation; with no real-run promises or strong marketing language.

```text
Treat model-compose as a pre-install experience asset, not an already-installed tool or a real runtime environment.

Output exactly four parts:
1. Ask me 3 necessary questions first.
2. Give an "experience script": use the three labels [Previewable before install], [Must verify after install], and [Insufficient evidence] to show how it might guide the workflow.
3. Give an after-install verification checklist: list which capabilities can only be confirmed after a real install, real host loading, and a real project run.
4. Give a cautious recommendation: only "worth researching/trialing further", "add information before deciding", or "not recommended to continue"; do not endorse the project.

Hard boundaries:
- Do not claim you have installed, run, executed tests, modified files, or produced real results.
- Do not write promise-like phrasing such as "auto-adapts", "guarantees passing", "perfect fit", or "strongly recommend installing".
- If you describe how it works after install, you must use a conditional such as "if installed successfully and the host loads the Skill correctly, it might...".
- The experience script may only be written as "example lines / hypothetical flow": use "might ask / might suggest / might show", not "has written, has generated, has passed, is running, is generating".
- Prompt Preview does not hand out install commands; if the user is ready to trial, only prompt them to read Quick Start and the Risk Card first and to verify in an isolated environment.
- Every project fact must come from a supported claim, evidence_refs, or source_paths; inferred/unverified items can only be risks or open questions.

```

### Role / Skill selection

- Goal: Pick the best-matching asset from the project's roles or Skills.
- Expected output: A list of candidate roles or Skills, each with an applicable scenario, evidence paths, risk boundary, and whether after-install verification is needed.

```text
Read role_skill_index and recommend 3-5 of the most relevant roles or Skills for my target task. For each recommendation, state the applicable scenario, likely output, risk boundary, and evidence_refs.
```

### Risk pre-check

- Goal: Identify environment, permission, rule-conflict, and quality risks before installing or adopting.
- Expected output: A checklist of environment, permission, dependency, license, host-conflict, quality risk, and unknown items.

```text
Based on risk_card, boundaries, and quick_start_candidates, give me a pre-install risk pre-check list. Do not run commands for me; only explain what I should check, why, and what impact a failure would have.
```

### Host AI kickoff instruction

- Goal: Turn the project context into a host AI instruction for the start of a conversation.
- Expected output: A pre-work instruction with clear boundaries and clear evidence citations, suitable to copy to a host AI.

```text
Based on the AI Context Pack for model-compose, generate a pre-work instruction I can paste to my host AI. This instruction must obey not_runtime=true and must not claim the project has been installed, run, or produced real results.
```

## Role / Skill Index

- Indexed 80 role / Skill / project-doc entries.

- **Model-Compose User Guide** (project_doc): Welcome to the model-compose user guide! This comprehensive documentation will help you master declarative AI workflow orchestration—from basic concepts to advanced deployment strategies. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `docs/user-guide/README.md`
- **Model-Compose 사용자 가이드** (project_doc): model-compose 사용자 가이드에 오신 것을 환영합니다! 이 포괄적인 문서는 기본 개념부터 고급 배포 전략까지 선언적 AI 워크플로우 오케스트레이션을 마스터하는 데 도움을 드립니다. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `docs/user-guide/ko/README.md`
- **Model-Compose 用户指南** (project_doc): 欢迎使用 model-compose 用户指南！本综合文档将帮助您掌握声明式 AI 工作流编排——从基础概念到高级部署策略。 Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `docs/user-guide/zh-cn/README.md`
- **model-compose** (project_doc): ! model-compose - Compose Any AI, Deploy Anywhere docs/images/main-banner.png Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `README.md`
- **Model-Compose Examples** (project_doc): This directory contains practical examples demonstrating various features and use cases of model-compose. Each example includes a ready-to-run model-compose.yml configuration file. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/README.md`
- **Code Reviewer Agent Example** (project_doc): This example demonstrates an autonomous agent that reads files, lists directories, and searches code to perform code reviews and provide improvement suggestions. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/agents/code-reviewer/README.md`
- **DESIGN.md Generator Example** (project_doc): This example demonstrates a declarative pipeline that analyzes a website's visual design system and generates a comprehensive DESIGN.md document, using headless browser automation with the web-browser component and specialized AI sub-agents powered by GPT-4o. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/agents/design-md-generator/README.md`
- **Disk Analyzer Agent Example** (project_doc): This example demonstrates an autonomous agent that uses shell commands as tools to analyze system disk usage and provide detailed recommendations. It is an agent-powered version of the analyze-disk-usage example. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/agents/disk-analyzer/README.md`
- **Human-in-the-Loop Agent Example** (project_doc): This example demonstrates a file management agent that requires human approval before executing dangerous operations write, delete , while allowing safe operations read, list to run without interruption. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/agents/human-in-the-loop/README.md`
- **K-POP Fancam Collector Example** (project_doc): This example demonstrates an autonomous agent workflow that searches YouTube for K-POP fancam videos from a natural-language prompt, and optionally filters them by orientation portrait/vertical . It combines a GPT-4o agent with two private tool workflows backed by the YouTube Data API and a lightweight web scraper. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/agents/kpop-fancam-collector/README.md`
- **Multi-Tool Assistant Agent Example** (project_doc): This example demonstrates a versatile assistant agent that combines multiple tools — web search, weather lookup, calculator, and clock — to answer a wide range of questions. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/agents/multi-tool/README.md`
- **RAG Assistant Agent Example** (project_doc): This example demonstrates an autonomous agent that uses Retrieval-Augmented Generation RAG to answer questions by searching and adding knowledge to a ChromaDB vector store. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/agents/rag-assistant/README.md`
- **Web Researcher Agent Example** (project_doc): This example demonstrates an autonomous agent that searches the web and fetches page content to research a topic and provide a comprehensive answer. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/agents/web-researcher/README.md`
- **Analyze Disk Usage Example** (project_doc): This example demonstrates a workflow that automatically analyzes system disk usage and provides detailed analysis using GPT-4o. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/analyze-disk-usage/README.md`
- **Audio Extractor Example** (project_doc): This example demonstrates an audio extractor using the audio-extractor component, showcasing how model-compose can orchestrate ffmpeg-based audio extraction from video or audio files with configurable encoding options. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/audio-extractor/README.md`
- **Telegram Bot** (project_doc): A Telegram bot that receives messages via webhook, processes them with OpenAI GPT-4o, and sends replies back to the user. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/channels/telegram/README.md`
- **Conditional Routing with if Example** (project_doc): Conditional Routing with if Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/conditional-routing/if/README.md`
- **Conditional Routing with random-router Example** (project_doc): Conditional Routing with random-router Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/conditional-routing/random/README.md`
- **Conditional Routing with switch Example** (project_doc): Conditional Routing with switch Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/conditional-routing/switch/README.md`
- **HuggingFace Datasets Example** (project_doc): This example demonstrates how to use model-compose with HuggingFace Datasets for loading, processing, and concatenating datasets from the HuggingFace Hub. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/datasets/huggingface/README.md`
- **Docker Nginx Example** (project_doc): This example demonstrates how to use Docker runtime for components, running an Nginx container that serves static files from a local directory with volume mounting. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/docker/README.md`
- **Echo Server Example** (project_doc): This example demonstrates a simple HTTP echo server that receives user input and returns it back, showcasing how model-compose can manage and communicate with local HTTP services. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/echo-server/README.md`
- **Cloudflare Named Tunnel Gateway Example** (project_doc): Cloudflare Named Tunnel Gateway Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/gateway/http-tunnel/cloudflare-named/README.md`
- **Cloudflare Quick Tunnel Gateway Example** (project_doc): Cloudflare Quick Tunnel Gateway Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/gateway/http-tunnel/cloudflare/README.md`
- **Ngrok HTTP Tunnel Gateway Example** (project_doc): This example demonstrates how to use ngrok HTTP tunnel gateway to expose local services to the internet. This enables external services to send callbacks to your local endpoints without requiring a public IP or SSH server. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/gateway/http-tunnel/ngrok/README.md`
- **SSH Tunnel Gateway Example** (project_doc): This example demonstrates how to use SSH tunnel gateway to expose local services to external networks through remote port forwarding. This enables external services to send callbacks to your local endpoints. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/gateway/ssh-tunnel/README.md`
- **ArangoDB Graph Store Example** (project_doc): This example demonstrates how to use model-compose with ArangoDB as a graph store for building and querying social graphs. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/graph-store/arangodb/README.md`
- **Neo4j Graph Store Example** (project_doc): This example demonstrates how to use model-compose with Neo4j as a graph store for building and querying knowledge graphs. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/graph-store/neo4j/README.md`
- **Image Processor Example** (project_doc): This example demonstrates a comprehensive image processing service using the image-processor component, showcasing how model-compose can orchestrate various image manipulation operations through a single component with multiple actions. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/image-processor/README.md`
- **Interrupt Example** (project_doc): This example demonstrates the Human-in-the-Loop HITL interrupt feature, which pauses a workflow for human review before and after executing a shell command. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/interrupt/README.md`
- **Redis Key-Value Store Example** (project_doc): This example demonstrates how to use model-compose with Redis as a key-value store for storing, retrieving, and managing data in workflows. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/key-value-store/redis/README.md`
- **Make Inspiring Quote Voice Example** (project_doc): This example demonstrates a complex multi-step workflow that combines text generation with speech synthesis, creating inspiring motivational quotes and converting them to natural-sounding audio using OpenAI GPT-4o and ElevenLabs Text-to-Speech. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/make-inspiring-quote-voice/README.md`
- **Korea DART MCP Server Example** (project_doc): This example demonstrates how to create an MCP server for querying Korean financial disclosure data from DART Data Analysis, Retrieval and Transfer System https://opendart.fss.or.kr , operated by South Korea's Financial Supervisory Service. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/mcp-servers/korea-dart-mcp/README.md`
- **Anthropic Chat Completions Stream Example** (project_doc): Anthropic Chat Completions Stream Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/anthropic/anthropic-chat-completions-stream/README.md`
- **Anthropic Chat Completions Example** (project_doc): This example demonstrates how to create a simple chat interface using Anthropic's Claude model through the Messages API. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/anthropic/anthropic-chat-completions/README.md`
- **ElevenLabs Text-to-Speech Example** (project_doc): This example demonstrates how to use model-compose with ElevenLabs AI to convert text into high-quality, natural-sounding speech. ElevenLabs provides state-of-the-art voice synthesis with multilingual support and realistic voice cloning capabilities. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/elevenlabs/elevenlabs-text-to-speech/README.md`
- **Google Cloud Vision API Example** (project_doc): This example demonstrates how to use Google Cloud Vision API for various image analysis tasks including label detection, text recognition OCR , face detection, object localization, landmark detection, and logo detection. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/google/google-cloud-vision/README.md`
- **OpenAI Text-to-Speech Example** (project_doc): This example demonstrates how to use model-compose with OpenAI's Text-to-Speech TTS API to convert text into natural-sounding speech using multiple high-quality voices and models. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/openai/openai-audio-speech/README.md`
- **OpenAI Audio Transcriptions Example** (project_doc): OpenAI Audio Transcriptions Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/openai/openai-audio-transciptions/README.md`
- **OpenAI Chat Completions Stream Example** (project_doc): OpenAI Chat Completions Stream Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/openai/openai-chat-completions-stream/README.md`
- **OpenAI Chat Completions Example** (project_doc): This example demonstrates how to create a simple chat interface using OpenAI's GPT-4o model through the Chat Completions API. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/openai/openai-chat-completions/README.md`
- **OpenAI Image Edits Example** (project_doc): This example demonstrates how to use model-compose with OpenAI's Image Editing API to modify images using text prompts and AI-powered image manipulation. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/openai/openai-image-edits/README.md`
- **OpenAI Image Generation Example** (project_doc): This example demonstrates how to generate images from text prompts using OpenAI's image generation models, including both DALL-E and GPT image models. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/openai/openai-image-generations/README.md`
- **OpenAI Image Variations Example** (project_doc): This example demonstrates how to use model-compose with OpenAI's Image Variations API to generate creative variations of existing images using DALL·E technology. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/openai/openai-image-variations/README.md`
- **xAI Chat Completion Example** (project_doc): This example demonstrates how to create a simple chat interface using xAI's Grok model through the Chat Completions API. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-providers/xai/xai-chat-completion/README.md`
- **Chat Completion Model Task Example** (project_doc): This example demonstrates how to use local language models for chat completion using model-compose's built-in chat-completion task with HuggingFace transformers, providing conversational AI capabilities without external API dependencies. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/chat-completion/huggingface/README.md`
- **Chat Completion llama.cpp Example** (project_doc): This example demonstrates how to run chat completion locally using GGUF format models with llama.cpp via model-compose's built-in llamacpp driver. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/chat-completion/llamacpp/README.md`
- **Image-to-Text Model Task Example** (project_doc): This example demonstrates how to use local vision-language models for image captioning and description using model-compose's built-in image-to-text task with HuggingFace transformers, providing offline image understanding capabilities. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/image-to-text/README.md`
- **Image Upscale Model Task Example** (project_doc): This example demonstrates how to use local super-resolution models for image upscaling using model-compose's built-in image-upscale task with Real-ESRGAN, providing offline image enhancement capabilities. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/image-upscale/README.md`
- **Music Generation Model Task Example** (project_doc): Music Generation Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/music-generation/README.md`
- **Speech-to-Text Model Task Example** (project_doc): This example demonstrates how to use a local Whisper model for audio transcription using model-compose's built-in speech-to-text task with HuggingFace transformers, providing offline speech recognition capabilities. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/speech-to-text/README.md`
- **Text Summarization Stream Model Task Example** (project_doc): Text Summarization Stream Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/summarization-stream/README.md`
- **Text Summarization Model Task Example** (project_doc): Text Summarization Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/summarization/README.md`
- **Text Classification Model Task Example** (project_doc): Text Classification Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-classification/README.md`
- **Text Embedding llama.cpp Example** (project_doc): This example demonstrates how to generate text embeddings locally using GGUF format embedding models with llama.cpp via model-compose's built-in llamacpp driver. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-embedding-llamacpp/README.md`
- **Text Embedding Model Task Example** (project_doc): This example demonstrates how to generate text embeddings using local sentence transformer models with model-compose's built-in text-embedding task, providing semantic vector representations of text for similarity search and ML applications. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-embedding/README.md`
- **Text Generation llama.cpp Example** (project_doc): This example demonstrates how to run text generation locally using GGUF format models with llama.cpp via model-compose's built-in llamacpp driver. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-generation-llamacpp/README.md`
- **Text Generation with Multiple LoRA Adapters** (project_doc): Text Generation with Multiple LoRA Adapters Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-generation-lora/README.md`
- **Text Generation Model Task Example** (project_doc): This example demonstrates how to use local language models for text generation using model-compose's built-in model task functionality with HuggingFace transformers. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-generation/README.md`
- **Text to Speech Voice Cloning Model Task Example** (project_doc): Text to Speech Voice Cloning Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-to-speech-clone/README.md`
- **Text to Speech Voice Design Model Task Example** (project_doc): Text to Speech Voice Design Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-to-speech-design/README.md`
- **Text to Speech Preset Voice Model Task Example** (project_doc): Text to Speech Preset Voice Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/text-to-speech-generate/README.md`
- **Text Translation Stream Model Task Example** (project_doc): Text Translation Stream Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/translation-stream/README.md`
- **Text Translation Model Task Example** (project_doc): Text Translation Model Task Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/model-tasks/translation/README.md`
- **SQLite Search Engine Example** (project_doc): This example demonstrates how to use model-compose with SQLite FTS5 as a full-text search engine for indexing, searching, and managing documents in workflows. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/search-engine/sqlite/README.md`
- **Text Splitter Example** (project_doc): This example demonstrates how to use model-compose with a text splitter component to break down large text documents into smaller, manageable chunks for AI processing and analysis. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/split-text/README.md`
- **ChromaDB Vector Store Example** (project_doc): This example demonstrates how to use model-compose with ChromaDB as a vector store for semantic search and similarity matching using text embeddings. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/vector-store/chroma/README.md`
- **Milvus Vector Store Example** (project_doc): This example demonstrates how to use model-compose with Milvus as a vector database for large-scale semantic search and similarity matching using text embeddings. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/vector-store/milvus/README.md`
- **VibeVoice Realtime TTS** (project_doc): Text-to-speech using Microsoft VibeVoice Realtime 0.5B https://github.com/microsoft/VibeVoice running in a Docker container via WebSocket. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/vibevoice-realtime-tts/README.md`
- **Video Converter Example** (project_doc): This example demonstrates a video format converter using the video-converter component, showcasing how model-compose can orchestrate ffmpeg-based video processing with configurable encoding options. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/video-converter/README.md`
- **Video Scene Detector Example** (project_doc): This example demonstrates how to use model-compose with the video-scene-detector component to detect scene changes in video files using different detection backends. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/video-scene-detector/README.md`
- **vLLM Chat Completion Stream Example** (project_doc): vLLM Chat Completion Stream Example Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/vllm-chat-completion-stream/README.md`
- **vLLM Text to Speech Example** (project_doc): This example demonstrates how to generate speech audio from text using the Qwen3-TTS model served via vLLM-Omni, with support for custom voice and multilingual synthesis. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/vllm-text-to-speech/README.md`
- **Web Browser Example** (project_doc): This example demonstrates headless browser automation using the web-browser component, with CAPTCHA detection and human-in-the-loop resolution via noVNC. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/web-browser/README.md`
- **Web Scraper Examples** (project_doc): This example demonstrates various web scraping capabilities using the web-scraper component with multiple workflows for different scraping scenarios. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/web-scraper/README.md`
- **Workflow Queue Stream Example** (project_doc): This example demonstrates how to stream workflow output across distributed instances using Redis. A dispatcher receives HTTP requests and forwards them to a remote subscriber, which calls the OpenAI streaming API and delivers chunks back through Redis Streams. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/workflow-queue-stream/README.md`
- **Workflow Queue Example** (project_doc): This example demonstrates how to distribute workflow execution across multiple instances using Redis as a message queue. A dispatcher receives requests and forwards them to a remote subscriber for processing. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `examples/workflow-queue/README.md`
- **Agent Component** (project_doc): The agent component enables building autonomous AI agents that use workflows as tools. It implements a ReAct Reasoning + Acting loop where an LLM iteratively reasons about a task, calls tools, and processes results until a final answer is produced. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `docs/reference/compose/components/agent.md`
- **Model Component** (project_doc): The model component enables loading and running AI/ML models locally using HuggingFace transformers. It supports various tasks including text generation, chat completion, text embedding, classification, translation, summarization, image-to-text processing, and text-to-speech synthesis. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `docs/reference/compose/components/model.md`
- **Chapter 1: Getting Started** (project_doc): model-compose is a declarative AI workflow orchestrator that lets you define and run AI model pipelines using simple YAML configuration files. Inspired by docker-compose , it brings the same philosophy of declarative configuration to the world of AI model orchestration. Activation hint: Reference this when the user needs to understand the project's structure, install path, or boundaries. Evidence: `docs/user-guide/01-getting-started.md`

## Evidence Index

- Indexed 80 evidence entries.

- **Model-Compose User Guide** (documentation): Welcome to the model-compose user guide! This comprehensive documentation will help you master declarative AI workflow orchestration—from basic concepts to advanced deployment strategies. Evidence: `docs/user-guide/README.md`
- **Model-Compose 사용자 가이드** (documentation): model-compose 사용자 가이드에 오신 것을 환영합니다! 이 포괄적인 문서는 기본 개념부터 고급 배포 전략까지 선언적 AI 워크플로우 오케스트레이션을 마스터하는 데 도움을 드립니다. Evidence: `docs/user-guide/ko/README.md`
- **Model-Compose 用户指南** (documentation): 欢迎使用 model-compose 用户指南！本综合文档将帮助您掌握声明式 AI 工作流编排——从基础概念到高级部署策略。 Evidence: `docs/user-guide/zh-cn/README.md`
- **model-compose** (documentation): ! model-compose - Compose Any AI, Deploy Anywhere docs/images/main-banner.png Evidence: `README.md`
- **Model-Compose Examples** (documentation): This directory contains practical examples demonstrating various features and use cases of model-compose. Each example includes a ready-to-run model-compose.yml configuration file. Evidence: `examples/README.md`
- **Code Reviewer Agent Example** (documentation): This example demonstrates an autonomous agent that reads files, lists directories, and searches code to perform code reviews and provide improvement suggestions. Evidence: `examples/agents/code-reviewer/README.md`
- **DESIGN.md Generator Example** (documentation): This example demonstrates a declarative pipeline that analyzes a website's visual design system and generates a comprehensive DESIGN.md document, using headless browser automation with the web-browser component and specialized AI sub-agents powered by GPT-4o. Evidence: `examples/agents/design-md-generator/README.md`
- **Disk Analyzer Agent Example** (documentation): This example demonstrates an autonomous agent that uses shell commands as tools to analyze system disk usage and provide detailed recommendations. It is an agent-powered version of the analyze-disk-usage example. Evidence: `examples/agents/disk-analyzer/README.md`
- **Human-in-the-Loop Agent Example** (documentation): This example demonstrates a file management agent that requires human approval before executing dangerous operations write, delete , while allowing safe operations read, list to run without interruption. Evidence: `examples/agents/human-in-the-loop/README.md`
- **K-POP Fancam Collector Example** (documentation): This example demonstrates an autonomous agent workflow that searches YouTube for K-POP fancam videos from a natural-language prompt, and optionally filters them by orientation portrait/vertical . It combines a GPT-4o agent with two private tool workflows backed by the YouTube Data API and a lightweight web scraper. Evidence: `examples/agents/kpop-fancam-collector/README.md`
- **Multi-Tool Assistant Agent Example** (documentation): This example demonstrates a versatile assistant agent that combines multiple tools — web search, weather lookup, calculator, and clock — to answer a wide range of questions. Evidence: `examples/agents/multi-tool/README.md`
- **RAG Assistant Agent Example** (documentation): This example demonstrates an autonomous agent that uses Retrieval-Augmented Generation RAG to answer questions by searching and adding knowledge to a ChromaDB vector store. Evidence: `examples/agents/rag-assistant/README.md`
- **Web Researcher Agent Example** (documentation): This example demonstrates an autonomous agent that searches the web and fetches page content to research a topic and provide a comprehensive answer. Evidence: `examples/agents/web-researcher/README.md`
- **Analyze Disk Usage Example** (documentation): This example demonstrates a workflow that automatically analyzes system disk usage and provides detailed analysis using GPT-4o. Evidence: `examples/analyze-disk-usage/README.md`
- **Audio Extractor Example** (documentation): This example demonstrates an audio extractor using the audio-extractor component, showcasing how model-compose can orchestrate ffmpeg-based audio extraction from video or audio files with configurable encoding options. Evidence: `examples/audio-extractor/README.md`
- **Telegram Bot** (documentation): A Telegram bot that receives messages via webhook, processes them with OpenAI GPT-4o, and sends replies back to the user. Evidence: `examples/channels/telegram/README.md`
- **Conditional Routing with if Example** (documentation): Conditional Routing with if Example Evidence: `examples/conditional-routing/if/README.md`
- **Conditional Routing with random-router Example** (documentation): Conditional Routing with random-router Example Evidence: `examples/conditional-routing/random/README.md`
- **Conditional Routing with switch Example** (documentation): Conditional Routing with switch Example Evidence: `examples/conditional-routing/switch/README.md`
- **HuggingFace Datasets Example** (documentation): This example demonstrates how to use model-compose with HuggingFace Datasets for loading, processing, and concatenating datasets from the HuggingFace Hub. Evidence: `examples/datasets/huggingface/README.md`
- **Docker Nginx Example** (documentation): This example demonstrates how to use Docker runtime for components, running an Nginx container that serves static files from a local directory with volume mounting. Evidence: `examples/docker/README.md`
- **Echo Server Example** (documentation): This example demonstrates a simple HTTP echo server that receives user input and returns it back, showcasing how model-compose can manage and communicate with local HTTP services. Evidence: `examples/echo-server/README.md`
- **Cloudflare Named Tunnel Gateway Example** (documentation): Cloudflare Named Tunnel Gateway Example Evidence: `examples/gateway/http-tunnel/cloudflare-named/README.md`
- **Cloudflare Quick Tunnel Gateway Example** (documentation): Cloudflare Quick Tunnel Gateway Example Evidence: `examples/gateway/http-tunnel/cloudflare/README.md`
- **Ngrok HTTP Tunnel Gateway Example** (documentation): This example demonstrates how to use ngrok HTTP tunnel gateway to expose local services to the internet. This enables external services to send callbacks to your local endpoints without requiring a public IP or SSH server. Evidence: `examples/gateway/http-tunnel/ngrok/README.md`
- **SSH Tunnel Gateway Example** (documentation): This example demonstrates how to use SSH tunnel gateway to expose local services to external networks through remote port forwarding. This enables external services to send callbacks to your local endpoints. Evidence: `examples/gateway/ssh-tunnel/README.md`
- **ArangoDB Graph Store Example** (documentation): This example demonstrates how to use model-compose with ArangoDB as a graph store for building and querying social graphs. Evidence: `examples/graph-store/arangodb/README.md`
- **Neo4j Graph Store Example** (documentation): This example demonstrates how to use model-compose with Neo4j as a graph store for building and querying knowledge graphs. Evidence: `examples/graph-store/neo4j/README.md`
- **Image Processor Example** (documentation): This example demonstrates a comprehensive image processing service using the image-processor component, showcasing how model-compose can orchestrate various image manipulation operations through a single component with multiple actions. Evidence: `examples/image-processor/README.md`
- **Interrupt Example** (documentation): This example demonstrates the Human-in-the-Loop HITL interrupt feature, which pauses a workflow for human review before and after executing a shell command. Evidence: `examples/interrupt/README.md`
- **Redis Key-Value Store Example** (documentation): This example demonstrates how to use model-compose with Redis as a key-value store for storing, retrieving, and managing data in workflows. Evidence: `examples/key-value-store/redis/README.md`
- **Make Inspiring Quote Voice Example** (documentation): This example demonstrates a complex multi-step workflow that combines text generation with speech synthesis, creating inspiring motivational quotes and converting them to natural-sounding audio using OpenAI GPT-4o and ElevenLabs Text-to-Speech. Evidence: `examples/make-inspiring-quote-voice/README.md`
- **Korea DART MCP Server Example** (documentation): This example demonstrates how to create an MCP server for querying Korean financial disclosure data from DART Data Analysis, Retrieval and Transfer System https://opendart.fss.or.kr , operated by South Korea's Financial Supervisory Service. Evidence: `examples/mcp-servers/korea-dart-mcp/README.md`
- **Anthropic Chat Completions Stream Example** (documentation): Anthropic Chat Completions Stream Example Evidence: `examples/model-providers/anthropic/anthropic-chat-completions-stream/README.md`
- **Anthropic Chat Completions Example** (documentation): This example demonstrates how to create a simple chat interface using Anthropic's Claude model through the Messages API. Evidence: `examples/model-providers/anthropic/anthropic-chat-completions/README.md`
- **ElevenLabs Text-to-Speech Example** (documentation): This example demonstrates how to use model-compose with ElevenLabs AI to convert text into high-quality, natural-sounding speech. ElevenLabs provides state-of-the-art voice synthesis with multilingual support and realistic voice cloning capabilities. Evidence: `examples/model-providers/elevenlabs/elevenlabs-text-to-speech/README.md`
- **Google Cloud Vision API Example** (documentation): This example demonstrates how to use Google Cloud Vision API for various image analysis tasks including label detection, text recognition OCR , face detection, object localization, landmark detection, and logo detection. Evidence: `examples/model-providers/google/google-cloud-vision/README.md`
- **OpenAI Text-to-Speech Example** (documentation): This example demonstrates how to use model-compose with OpenAI's Text-to-Speech TTS API to convert text into natural-sounding speech using multiple high-quality voices and models. Evidence: `examples/model-providers/openai/openai-audio-speech/README.md`
- **OpenAI Audio Transcriptions Example** (documentation): OpenAI Audio Transcriptions Example Evidence: `examples/model-providers/openai/openai-audio-transciptions/README.md`
- **OpenAI Chat Completions Stream Example** (documentation): OpenAI Chat Completions Stream Example Evidence: `examples/model-providers/openai/openai-chat-completions-stream/README.md`
- **OpenAI Chat Completions Example** (documentation): This example demonstrates how to create a simple chat interface using OpenAI's GPT-4o model through the Chat Completions API. Evidence: `examples/model-providers/openai/openai-chat-completions/README.md`
- **OpenAI Image Edits Example** (documentation): This example demonstrates how to use model-compose with OpenAI's Image Editing API to modify images using text prompts and AI-powered image manipulation. Evidence: `examples/model-providers/openai/openai-image-edits/README.md`
- **OpenAI Image Generation Example** (documentation): This example demonstrates how to generate images from text prompts using OpenAI's image generation models, including both DALL-E and GPT image models. Evidence: `examples/model-providers/openai/openai-image-generations/README.md`
- **OpenAI Image Variations Example** (documentation): This example demonstrates how to use model-compose with OpenAI's Image Variations API to generate creative variations of existing images using DALL·E technology. Evidence: `examples/model-providers/openai/openai-image-variations/README.md`
- **xAI Chat Completion Example** (documentation): This example demonstrates how to create a simple chat interface using xAI's Grok model through the Chat Completions API. Evidence: `examples/model-providers/xai/xai-chat-completion/README.md`
- **Chat Completion Model Task Example** (documentation): This example demonstrates how to use local language models for chat completion using model-compose's built-in chat-completion task with HuggingFace transformers, providing conversational AI capabilities without external API dependencies. Evidence: `examples/model-tasks/chat-completion/huggingface/README.md`
- **Chat Completion llama.cpp Example** (documentation): This example demonstrates how to run chat completion locally using GGUF format models with llama.cpp via model-compose's built-in llamacpp driver. Evidence: `examples/model-tasks/chat-completion/llamacpp/README.md`
- **Image-to-Text Model Task Example** (documentation): This example demonstrates how to use local vision-language models for image captioning and description using model-compose's built-in image-to-text task with HuggingFace transformers, providing offline image understanding capabilities. Evidence: `examples/model-tasks/image-to-text/README.md`
- **Image Upscale Model Task Example** (documentation): This example demonstrates how to use local super-resolution models for image upscaling using model-compose's built-in image-upscale task with Real-ESRGAN, providing offline image enhancement capabilities. Evidence: `examples/model-tasks/image-upscale/README.md`
- **Music Generation Model Task Example** (documentation): Music Generation Model Task Example Evidence: `examples/model-tasks/music-generation/README.md`
- **Speech-to-Text Model Task Example** (documentation): This example demonstrates how to use a local Whisper model for audio transcription using model-compose's built-in speech-to-text task with HuggingFace transformers, providing offline speech recognition capabilities. Evidence: `examples/model-tasks/speech-to-text/README.md`
- **Text Summarization Stream Model Task Example** (documentation): Text Summarization Stream Model Task Example Evidence: `examples/model-tasks/summarization-stream/README.md`
- **Text Summarization Model Task Example** (documentation): Text Summarization Model Task Example Evidence: `examples/model-tasks/summarization/README.md`
- **Text Classification Model Task Example** (documentation): Text Classification Model Task Example Evidence: `examples/model-tasks/text-classification/README.md`
- **Text Embedding llama.cpp Example** (documentation): This example demonstrates how to generate text embeddings locally using GGUF format embedding models with llama.cpp via model-compose's built-in llamacpp driver. Evidence: `examples/model-tasks/text-embedding-llamacpp/README.md`
- **Text Embedding Model Task Example** (documentation): This example demonstrates how to generate text embeddings using local sentence transformer models with model-compose's built-in text-embedding task, providing semantic vector representations of text for similarity search and ML applications. Evidence: `examples/model-tasks/text-embedding/README.md`
- **Text Generation llama.cpp Example** (documentation): This example demonstrates how to run text generation locally using GGUF format models with llama.cpp via model-compose's built-in llamacpp driver. Evidence: `examples/model-tasks/text-generation-llamacpp/README.md`
- **Text Generation with Multiple LoRA Adapters** (documentation): Text Generation with Multiple LoRA Adapters Evidence: `examples/model-tasks/text-generation-lora/README.md`
- **Text Generation Model Task Example** (documentation): This example demonstrates how to use local language models for text generation using model-compose's built-in model task functionality with HuggingFace transformers. Evidence: `examples/model-tasks/text-generation/README.md`
- **Text to Speech Voice Cloning Model Task Example** (documentation): Text to Speech Voice Cloning Model Task Example Evidence: `examples/model-tasks/text-to-speech-clone/README.md`
- The remaining 20 evidence entries are in `AI_CONTEXT_PACK.json` or `EVIDENCE_INDEX.json`.

## Rules the Host AI Must Follow

- **Treat this asset as pre-work context, not a runtime environment.**: The AI Context Pack contains only an evidence-backed understanding of the project, not the project's executable state. Evidence: `docs/user-guide/README.md`, `docs/user-guide/ko/README.md`, `docs/user-guide/zh-cn/README.md`
- **When answering the user, distinguish what can be previewed from what can only be verified after install.**: The consumer value of the pre-install experience comes from reducing bad installs and misjudgments, not from pretending to be a real run. Evidence: `docs/user-guide/README.md`, `docs/user-guide/ko/README.md`, `docs/user-guide/zh-cn/README.md`

## Questions the User Should Answer First

- Which host AI or local environment do you plan to use it in?
- Do you just want to experience the workflow first, or are you ready to actually install?
- What matters most to you: install cost, output quality, or conflicts with your existing rules?

## Acceptance Checks

- Every capability claim can be traced back to a file path in evidence_refs.
- AI_CONTEXT_PACK.md does not package previews as a real run.
- The user can understand who it fits, what it can do, how to start, and the risk boundaries within 3 minutes.

---

## Doramagic Context Augmentation

The following sections strengthen the repository context for a host AI. Human Manual data is a reading route, and pitfall notes become operating constraints.

## Human Manual Outline

Usage rule: this is only a reading route and salience signal, not factual authority. Concrete claims must still return to repo evidence or Claim Graph.

Host AI hard rules:
- Do not treat page titles, section order, summaries, or importance values as factual project evidence.
- When explaining the Human Manual outline, state that it is only a reading route or salience signal.
- Capability, installation, compatibility, runtime state, and risk claims must cite repo evidence, source paths, or Claim Graph.

- **Introduction and Core Philosophy**: importance `high`
  - source_paths: README.md, pyproject.toml, src/mindor/__init__.py, src/mindor/dsl/schema/compose.py, docs/user-guide/02-core-concepts.md
- **Component System and Architecture**: importance `high`
  - source_paths: src/mindor/core/component/component.py, src/mindor/dsl/schema/component/component.py, src/mindor/core/component/runtime/docker.py, src/mindor/core/component/services/agent.py, docs/reference/compose/components/model.md
- **Workflow Composition, Jobs and Streaming**: importance `high`
  - source_paths: src/mindor/core/workflow/workflow.py, src/mindor/core/workflow/job/job.py, src/mindor/core/workflow/job/impl/if_.py, src/mindor/core/workflow/job/impl/for_each.py, src/mindor/core/workflow/job/impl/switch.py
- **Deployment, Runtimes, and Protocol Adapters**: importance `high`
  - source_paths: src/mindor/core/controller/controller.py, src/mindor/core/runtime/docker.py, src/mindor/core/runtime/native.py, src/mindor/core/controller/adapters/services/http_server.py, src/mindor/core/controller/adapters/services/mcp_server.py

## Repo Inspection Evidence

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `1f7d7c4a834e4eb45491aa8d619a51842eb13fd3`
- inspected_files: `README.md`, `pyproject.toml`, `docs/reference/cli.md`, `docs/reference/compose/component.md`, `docs/reference/compose/components/agent.md`, `docs/reference/compose/components/datasets.md`, `docs/reference/compose/components/file-store.md`, `docs/reference/compose/components/graph-store.md`, `docs/reference/compose/components/http-client.md`, `docs/reference/compose/components/http-server.md`, `docs/reference/compose/components/image-processor.md`, `docs/reference/compose/components/key-value-store.md`, `docs/reference/compose/components/mcp-client.md`, `docs/reference/compose/components/mcp-server.md`, `docs/reference/compose/components/model-memory.md`, `docs/reference/compose/components/model-tokenizer.md`, `docs/reference/compose/components/model-trainer.md`, `docs/reference/compose/components/model.md`, `docs/reference/compose/components/search-engine.md`, `docs/reference/compose/components/shell.md`

Host AI hard rules:
- Without repo_clone_verified=true, do not claim that the source code has been read.
- Without repo_inspection_verified=true, do not write README, docs, or package-file conclusions as facts.
- Without quick_start_verified=true, do not claim that the Quick Start path has run successfully.

## Doramagic Pitfall Constraints

These rules come from Doramagic discovery, validation, or compilation findings. The host AI must treat them as operating constraints, not background notes.

### Constraint 1: Capability evidence risk requires verification

- Trigger: README/documentation is current enough for a first validation pass.
- Host AI rule: Reproduce the official install and quickstart path in an isolated environment.
- Why it matters: May increase setup, validation, or first-run risk for the user.
- Evidence: capability.assumptions | https://github.com/hanyeol/model-compose
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 2: Security or permission risk requires verification

- Trigger: no_demo
- Host AI rule: Reproduce the official install and quickstart path in an isolated environment.
- Why it matters: May increase setup, validation, or first-run risk for the user.
- Evidence: downstream_validation.risk_items | https://github.com/hanyeol/model-compose
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 3: Security or permission risk requires verification

- Trigger: no_demo
- Host AI rule: Reproduce the official install and quickstart path in an isolated environment.
- Why it matters: May increase setup, validation, or first-run risk for the user.
- Evidence: risks.scoring_risks | https://github.com/hanyeol/model-compose
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 4: Maintenance risk requires verification

- Trigger: issue_or_pr_quality=unknown。
- Host AI rule: Reproduce the official install and quickstart path in an isolated environment.
- Why it matters: May increase setup, validation, or first-run risk for the user.
- Evidence: evidence.maintainer_signals | https://github.com/hanyeol/model-compose
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 5: Maintenance risk requires verification

- Trigger: release_recency=unknown。
- Host AI rule: Reproduce the official install and quickstart path in an isolated environment.
- Why it matters: May increase setup, validation, or first-run risk for the user.
- Evidence: evidence.maintainer_signals | https://github.com/hanyeol/model-compose
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.