skills.rest
Military-Veteran-Team-LPT-Realty
Military-Veteran-Team-LPT-RealtyOFFICIAL
Explore all skills by this author →
GitHub

webdev-llm-integration

Integrates built-in LLM chat completions, structured JSON, and reasoning into Manus webdev projects.

Updated Sep 16, 2026
Search Tags:#llm#streaming#reasoning#trpc#json schema#chat completions#expo
One-click install
npx skills add https://github.com/Military-Veteran-Team-LPT-Realty/mvt-manus-public-skills --skill webdev-llm-integration-military-veteran-team-lpt-realty
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: webdev-llm-integration
Source: https://github.com/Military-Veteran-Team-LPT-Realty/mvt-manus-public-skills/tree/main/skills/webdev-llm-integration
Command: npx skills add https://github.com/Military-Veteran-Team-LPT-Realty/mvt-manus-public-skills --skill webdev-llm-integration-military-veteran-team-lpt-realty

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires streamdown.

What problem does it solve? Adding AI features to a Manus fullstack (web-db-user) or Expo mobile project normally requires managing API keys, model endpoints, and response parsing. This Skill provides preconfigured server-side LLM helpers with platform-injected credentials, so you can call chat completions, structured JSON outputs, and reasoning modes without manual setup. ## Core Features & Use Cases - Chat Completions via invokeLLM: Send system/user/assistant messages with text, image, or file content from server-side code such as tRPC procedures. - Structured JSON Responses: Enforce JSON Schema output through response_format for reliable data extraction tasks. - Model Discovery & Reasoning: List available models at runtime with listLLMModels and pass thinking/reasoning parameters for Claude, GPT-5, and Gemini families. - Use Case: Build an AI feature that extracts a name and age from user-submitted text by calling invokeLLM with a strict json_schema, then render the markdown response in the UI with Streamdown. ## Quick Start Ask the AI to add a server-side tRPC procedure that calls invokeLLM with a user prompt and returns the model's response to the frontend.

Frequently Asked Questions about webdev-llm-integration

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

FAQPage Schema
How do I call an LLM from a Manus webdev project?▼

Import invokeLLM from ./server/_core/llm and call it with a messages array containing system and user roles. Always invoke it from server-side code such as tRPC procedures so credentials stay on the server.

How to get structured JSON output from an LLM chat completion?▼

Pass a response_format object with type json_schema and a strict schema defining properties, required fields, and additionalProperties false. The model returns JSON matching the schema, accessible via choices[0].message.content.

Does invokeLLM support streaming responses?▼

All models support streaming, but invokeLLM does not expose a stream parameter by default. You can modify the helper to pass stream true and parse the SSE response, listening on res close with a finished flag to avoid aborting upstream.

How do I enable thinking or reasoning mode for Claude or GPT-5?▼

Pass model-specific parameters: thinking with type enabled and budget_tokens for Claude and Gemini, or reasoning with an effort level for GPT-5. Check capabilities.thinking_example from the models catalog for exact shapes.

Can I list available LLM models at runtime instead of hardcoding IDs?▼

Yes, call listLLMModels to get OpenAI-standard model metadata, then filter the returned IDs and pass the chosen one to invokeLLM. This avoids hardcoding model names that may change.

Related Skills

Explore More Software Engineering Skills→
TekLabsVNTekLabsVN
1

web-to-mobile

Convert an existing web project into an Expo mobile app with wired authentication.

Official
Advanced
TKlerxTKlerx

streamdown

Render streaming Markdown in React with syntax highlighting, Mermaid diagrams, and math.

Community
Advanced
kngender5kngender5

colab-multimodal-pipelines

Combine text, image, audio, and video processing in Colab with vision LLMs, TTS, ASR, and image generation.

Community
Advanced
tharunramagiritharunramagiri

onecli-gateway

Inject managed credentials into outbound HTTPS API calls.

Community
Advanced