krisspy-ai

v1.1.53

Published

9 hours ago

Unified AI Agent Library - Supports Claude Agent SDK, Codex SDK, and proxy providers (OpenAI, Gemini, Z.AI)

0High
0Medium
0Low

sofiane_krisspy

ai claude codex openai gemini anthropic agent sdk proxy

krisspy-ai

Official library for building AI agents on the Krisspy platform.

What is krisspy-ai?

krisspy-ai enables you to create custom agents using the Claude Agent SDK, while making it compatible with many models through an integrated proxy system.

Key Features

Multi-providers: Anthropic, OpenAI, Gemini, AWS Bedrock, Azure AI, Z.AI
Advanced reasoning: Extended thinking supported across all providers
Code execution: Agents can execute code in real-time
Sub-agents: Build multi-agent architectures
Skills: Use Anthropic Skills (Office files, etc.)
MCP: Connect external tools via MCP protocol
GenAI Services: Image generation, video generation, speech-to-text, text-to-speech

All of this without worrying about the underlying model - the proxy automatically translates requests.

Installation

npm install krisspy-ai

Quick Start

With Anthropic (direct)

import { query } from 'krisspy-ai';

for await (const event of query({
  prompt: 'Analyze this code and suggest improvements',
  options: {
    provider: 'anthropic',
    apiKey: 'sk-ant-...',
    model: 'sonnet'  // haiku, sonnet, opus
  }
})) {
  console.log(event);
}

With AWS Bedrock

import { query } from 'krisspy-ai';

for await (const event of query({
  prompt: 'Create an optimized sorting function',
  options: {
    provider: 'bedrock',
    accessKeyId: 'AKIA...',
    secretAccessKey: '...',
    region: 'us-west-2',
    model: 'sonnet'  // haiku, haiku-4.5, sonnet, sonnet-4, sonnet-4.5, opus, opus-4, opus-4.5
  }
})) {
  console.log(event);
}

With OpenAI (via proxy)

import { query } from 'krisspy-ai';

for await (const event of query({
  prompt: 'Explain design patterns to me',
  options: {
    provider: 'openai',
    apiKey: 'sk-...',
    model: 'gpt-4o'
  }
})) {
  console.log(event);
}

Supported Providers

| Provider | Description | Models | |----------|-------------|--------| | anthropic | Direct Anthropic API | haiku, sonnet, opus | | bedrock | AWS Bedrock (Claude) | haiku, sonnet, opus + 4/4.5 versions | | azure | Azure AI Foundry | gpt-4o, gpt-5.2 | | openai | OpenAI via proxy | gpt-4o, gpt-5.2, o1, o3 | | gemini | Google Gemini via proxy | gemini-2.5-pro, gemini-2.5-flash | | zai | Z.AI via proxy | glm-4.7 | | zai_direct | Z.AI Anthropic endpoint | haiku, sonnet, opus |

Bedrock Models

| Alias | Bedrock Model ID | |-------|------------------| | haiku | global.anthropic.claude-haiku-4-5-20251001-v1:0 | | haiku-4.5 | global.anthropic.claude-haiku-4-5-20251001-v1:0 | | sonnet | global.anthropic.claude-sonnet-4-20250514-v1:0 | | sonnet-4 | global.anthropic.claude-sonnet-4-20250514-v1:0 | | sonnet-4.5 | us.anthropic.claude-sonnet-4-5-20250929-v1:0 | | opus | global.anthropic.claude-opus-4-5-20251101-v1:0 | | opus-4 | global.anthropic.claude-opus-4-20250514-v1:0 | | opus-4.5 | global.anthropic.claude-opus-4-5-20251101-v1:0 |

GenAI Services

krisspy-ai includes generative AI services for image, video, and audio. These services work with both OpenAI and Azure OpenAI.

Image Generation

Generate images using GPT-Image-1 (OpenAI) or Azure DALL-E/GPT-Image.

import { generateImage } from 'krisspy-ai';

// With OpenAI
const result = await generateImage({
  service: 'openai',
  apiKey: 'sk-...',
  deploymentName: 'gpt-image-1',  // or 'dall-e-3'
  prompt: 'A cute baby polar bear playing in the snow',
  size: '1024x1024',  // '1024x1024', '1792x1024', '1024x1792'
});

console.log(result.data[0].url);

// With Azure OpenAI
const azureResult = await generateImage({
  service: 'azure',
  apiKey: 'your-azure-key',
  baseUrl: 'https://your-resource.openai.azure.com',
  deploymentName: 'gpt-image-1',
  apiVersion: '2024-02-15-preview',
  prompt: 'A futuristic city at sunset',
  size: '1792x1024',
  quality: 'hd',
});

Video Generation (Sora)

Generate videos using OpenAI Sora or Azure Sora.

import { generateVideo } from 'krisspy-ai';

// Async generator with status updates
for await (const event of generateVideo({
  service: 'openai',
  apiKey: 'sk-...',
  deploymentName: 'sora',
  prompt: 'Woolly mammoths walking through a snowy tundra',
  duration: 5,  // seconds
  width: 1920,
  height: 1080,
})) {
  if (event.type === 'status') {
    console.log(`Status: ${event.status.status}, Progress: ${event.status.progress}%`);
  } else if (event.type === 'result') {
    console.log(`Video ready: ${event.data.generations[0].url}`);
  }
}

// With Azure OpenAI
for await (const event of generateVideo({
  service: 'azure',
  apiKey: 'your-azure-key',
  baseUrl: 'https://your-resource.openai.azure.com',
  deploymentName: 'sora-2',
  prompt: 'A drone flying over mountains at sunrise',
  duration: 10,
})) {
  // Same event handling
}

Speech-to-Text (Transcription)

Transcribe audio using Whisper or GPT-4o-transcribe.

import { transcribe } from 'krisspy-ai';
import fs from 'fs';

const audioBuffer = fs.readFileSync('audio.mp3');

// With OpenAI
const result = await transcribe({
  service: 'openai',
  apiKey: 'sk-...',
  deploymentName: 'whisper-1',  // or 'gpt-4o-transcribe'
  audio: audioBuffer,
  language: 'en',
  responseFormat: 'verbose_json',
  timestampGranularities: ['word', 'segment'],
});

console.log(result.text);
console.log(result.words);     // Word-level timestamps
console.log(result.segments);  // Segment-level timestamps

// With Azure OpenAI
const azureResult = await transcribe({
  service: 'azure',
  apiKey: 'your-azure-key',
  baseUrl: 'https://your-resource.openai.azure.com',
  deploymentName: 'whisper',
  apiVersion: '2025-03-01-preview',
  audio: audioBuffer,
});

// With speaker diarization (Azure gpt-4o-transcribe-diarize)
const diarizedResult = await transcribe({
  service: 'azure',
  apiKey: 'your-azure-key',
  baseUrl: 'https://your-resource.openai.azure.com',
  deploymentName: 'gpt-4o-transcribe-diarize',
  apiVersion: '2025-03-01-preview',
  audio: audioBuffer,
});

console.log(diarizedResult.utterances);  // Speaker-separated segments

Text-to-Speech (TTS)

Convert text to speech using TTS-1, TTS-1-HD, or GPT-4o-mini-tts.

import { synthesize } from 'krisspy-ai';
import fs from 'fs';

// With OpenAI
const result = await synthesize({
  service: 'openai',
  apiKey: 'sk-...',
  deploymentName: 'tts-1',  // 'tts-1', 'tts-1-hd', 'gpt-4o-mini-tts'
  input: 'Hello! Welcome to krisspy-ai.',
  voice: 'nova',  // alloy, ash, coral, echo, fable, nova, onyx, sage, shimmer
  speed: 1.0,     // 0.25 to 4.0
  responseFormat: 'mp3',
});

fs.writeFileSync('output.mp3', result.audio);

// With Azure OpenAI (only supports: alloy, echo, fable, nova, onyx, shimmer)
const azureResult = await synthesize({
  service: 'azure',
  apiKey: 'your-azure-key',
  baseUrl: 'https://your-resource.openai.azure.com',
  deploymentName: 'tts',
  apiVersion: '2025-03-01-preview',
  input: 'Hello from Azure!',
  voice: 'nova',
});

// With voice styling (gpt-4o-mini-tts only)
const styledResult = await synthesize({
  service: 'openai',
  apiKey: 'sk-...',
  deploymentName: 'gpt-4o-mini-tts',
  input: 'This is exciting news!',
  voice: 'coral',
  instructions: 'Speak with enthusiasm and energy',
});

GenAI Service Options

Image Generation Options

| Option | Type | Description | |--------|------|-------------| | service | 'openai' \| 'azure' | Provider to use | | apiKey | string | API key | | baseUrl | string | Base URL (required for Azure) | | deploymentName | string | Model/deployment name | | prompt | string | Image description | | size | string | '1024x1024', '1792x1024', '1024x1792' | | quality | string | 'standard' or 'hd' | | n | number | Number of images (1-10) |

Video Generation Options

| Option | Type | Description | |--------|------|-------------| | service | 'openai' \| 'azure' | Provider to use | | apiKey | string | API key | | baseUrl | string | Base URL (required for Azure) | | deploymentName | string | Model/deployment name | | prompt | string | Video description | | duration | number | Duration in seconds | | width | number | Video width (default: 1920) | | height | number | Video height (default: 1080) | | nVariants | number | Number of variants |

Transcription Options

| Option | Type | Description | |--------|------|-------------| | service | 'openai' \| 'azure' | Provider to use | | apiKey | string | API key | | baseUrl | string | Base URL (required for Azure) | | deploymentName | string | Model/deployment name | | audio | Buffer \| string | Audio data or file path | | language | string | ISO-639-1 language code | | responseFormat | string | 'json', 'text', 'srt', 'vtt', 'verbose_json' | | timestampGranularities | string[] | ['word'], ['segment'], or both |

TTS Options

| Option | Type | Description | |--------|------|-------------| | service | 'openai' \| 'azure' | Provider to use | | apiKey | string | API key | | baseUrl | string | Base URL (required for Azure) | | deploymentName | string | Model/deployment name | | input | string | Text to synthesize (max 4096 chars) | | voice | string | Voice name | | speed | number | Speed (0.25 to 4.0) | | responseFormat | string | 'mp3', 'opus', 'aac', 'flac', 'wav', 'pcm' | | instructions | string | Voice styling (gpt-4o-mini-tts only) |

Advanced Options

const options = {
  // Provider
  provider: 'anthropic',
  apiKey: 'sk-ant-...',
  model: 'sonnet',
  baseUrl: 'https://api.anthropic.com',

  // Extended Thinking
  maxThinkingTokens: 10000,

  // MCP Servers
  mcpServers: {
    filesystem: {
      command: 'npx',
      args: ['-y', '@anthropic-ai/mcp-server-filesystem', '/path']
    },
    http_server: {
      type: 'http',
      url: 'http://localhost:3001/mcp'
    }
  },

  // Skills (Office files)
  betas: ['skills-2025-10-02'],

  // Attachments
  attachments: {
    images: [{ type: 'url', url: 'https://...' }],
    files: [{ type: 'pdf', media_type: 'application/pdf', data: 'base64...' }]
  },

  // Execution limits
  maxTurns: 50,
  maxBudgetUsd: 5.00,

  // Tools
  allowedTools: ['Read', 'Edit', 'Bash', 'WebSearch'],
  permissionMode: 'acceptEdits',

  // AWS Bedrock
  accessKeyId: 'AKIA...',
  secretAccessKey: '...',
  region: 'us-west-2',

  // Azure AI
  deploymentName: 'my-deployment',
  apiVersion: '2024-02-15-preview'
};

Architecture

┌─────────────────────────────────────────────────────────────┐
│                        krisspy-ai                           │
├─────────────────────────────────────────────────────────────┤
│                                                             │
│  ┌──────────────┐    ┌──────────────┐    ┌──────────────┐  │
│  │   Anthropic  │    │   Bedrock    │    │    Azure     │  │
│  │   (direct)   │    │   (native)   │    │   (proxy)    │  │
│  └──────────────┘    └──────────────┘    └──────────────┘  │
│                                                             │
│  ┌──────────────┐    ┌──────────────┐    ┌──────────────┐  │
│  │   OpenAI     │    │   Gemini     │    │    Z.AI      │  │
│  │   (proxy)    │    │   (proxy)    │    │   (proxy)    │  │
│  └──────────────┘    └──────────────┘    └──────────────┘  │
│                                                             │
│                    ┌──────────────────┐                    │
│                    │  Claude Agent    │                    │
│                    │      SDK         │                    │
│                    └──────────────────┘                    │
│                            │                               │
│         ┌──────────────────┼──────────────────┐           │
│         ▼                  ▼                  ▼           │
│  ┌────────────┐    ┌────────────┐    ┌────────────┐      │
│  │ Reasoning  │    │   Code     │    │    MCP     │      │
│  │            │    │ Execution  │    │   Tools    │      │
│  └────────────┘    └────────────┘    └────────────┘      │
│                                                             │
│  ┌─────────────────────────────────────────────────────┐  │
│  │              GenAI Services                          │  │
│  │  ┌─────────┐  ┌─────────┐  ┌─────────┐  ┌────────┐  │  │
│  │  │  Image  │  │  Video  │  │   STT   │  │  TTS   │  │  │
│  │  │ (DALL-E)│  │ (Sora)  │  │(Whisper)│  │        │  │  │
│  │  └─────────┘  └─────────┘  └─────────┘  └────────┘  │  │
│  │         OpenAI  |  Azure OpenAI                      │  │
│  └─────────────────────────────────────────────────────┘  │
│                                                             │
└─────────────────────────────────────────────────────────────┘

License

ISC

Published

Vulnerabilities

Links

Maintainers

Keywords

Readme

krisspy-ai

What is krisspy-ai?

Key Features

Installation

Quick Start

With Anthropic (direct)

With AWS Bedrock

With OpenAI (via proxy)

Supported Providers

Bedrock Models

GenAI Services

Image Generation

Video Generation (Sora)

Speech-to-Text (Transcription)

Text-to-Speech (TTS)

GenAI Service Options

Image Generation Options

Video Generation Options

Transcription Options

TTS Options

Advanced Options

Architecture

License

Links