mcp-ai mcp

v1.0.0 · MCP Tool · ai · registry.pascalai.org

Generic AI inference abstraction: chat, text completion, vision and tool-use across multiple providers (OpenAI, Anthropic, Cohere, etc.).

aillmchatgptclaudeinference

Input Parameters

Parameter	Type	Description
providerrequired	string	One of: `openai`, `anthropic`, `cohere`, `mistral`, `groq`, `together`, `perplexity`, `ollama`. Default: `openai`.
api_keyrequired	string
modelrequired	string
operationrequired	string	One of: `chat`, `complete`, `embed`, `vision`.
messagesoptional	array[object]
promptoptional	string
systemoptional	string
image_urloptional	string
max_tokensoptional	integer	Default: `1024`.
temperatureoptional	number	Default: `0.7`.
streamoptional	boolean	Default: `False`.

Output Fields

Field	Type	Description
content	string
model	string
provider	string
usage	object
embeddings	array[number]

Examples

Chat with any LLM

// Input
{
  "provider": "openai",
  "api_key": "sk-xxx",
  "model": "gpt-4o",
  "operation": "chat",
  "messages": [
    {
      "role": "user",
      "content": "Summarize the key benefits of microservices."
    }
  ]
}

// Output
{
  "content": "Microservices offer...",
  "model": "gpt-4o"
}

Install & Discovery

Install

ppm install mcp-ai

Get JSON Schema

GET /v1/packages/mcp-ai/1.0.0/schema

Discover by keyword

GET /v1/mcp/discover?q=ai

Discovery hint: Install with ppm install mcp-ai or invoke remotely via POST /v1/invoke/mcp-ai on the MCP Service.

PascalAI Usage

uses toolslib;
var Tool := LoadTool('mcp-ai');
var R := Tool.Call(JsonObj([]));
Writeln(R.ToJSON);