README.md

@fabrice-ai/tools

📅 Get Current Date

packages/tools/src/date.ts

Retrieves the current date in ISO format.

Example

import { getCurrentDate } from '@fabrice-ai/tools/date'

const exampleAgent = agent({
  role: '...',
  description: '...',
  tools: {
    getCurrentDate,
  },
})

🔍 Google Search with Serply.io

packages/tools/src/webSearch.ts

Performs a Google search with a search query using the Serply API.

Example

import { createWebSearchTools } from '@fabrice-ai/tools/webSearch'

const apiKey = 'your-serply-api-key'
const { googleSearch } = createWebSearchTools({ apiKey })

const exampleAgent = agent({
  role: '...',
  description: '...',
  tools: {
    googleSearch,
  },
})

🖼️ Google Image Search

packages/tools/src/webSearch.ts

Performs a Google Image search with a search query using the Serply API.

Example

import { createWebSearchTools } from '@fabrice-ai/tools/webSearch'

const apiKey = 'your-serply-api-key'
const { googleImageSearch } = createWebSearchTools({ apiKey })

const exampleAgent = agent({
  role: '...',
  description: '...',
  tools: {
    googleImageSearch,
  },
})

🌐 Web Scraper

packages/tools/src/firecrawl.ts

Scrapes a website using Firecrawl.dev to extract metadata and content.

Example

import { createFireCrawlTool } from '@fabrice-ai/tools/firecrawl'

const exampleAgent = agent({
  role: '...',
  description: '...',
  tools: {
    createFireCrawlTool({
        apiKey: 'Firecrawl API key ...'
    }),
  },
})

🫙 Vector Store

packages/tools/src/vector.ts

Save documents with embeddings, perform vector search.

Example

import { createVectorStoreTools } from '@fabrice-ai/tools/vector'

const { saveDocumentInVectorStore, searchInVectorStore } = createVectorStoreTools()

const exampleAgent = agent({
  role: '...',
  description: '...',
  tools: {
    saveDocumentInVectorStore,
    searchInVectorStore
  },
})

Example with custom VectorStore

import { createVectorStoreTools } from '@fabrice-ai/tools/vector'

/**
 * createVectorStore accepts a `vectorStore` adapter. 
 * This is a way to switch the default - in-memory store to Pinecone or others of your choice.
 */
const createPineconeVectorStore = () => {
  const store = new Map<string, EmbeddingResult>()

  const set = async (id: string, value: EmbeddingResult): Promise<void> => {
    // @tbd: implement storing document in Pinecone
  }

  const entries = async (): Promise<[string, EmbeddingResult][]> => {
    // @tbd: implement searching documents in Pinecone
  }

  return {
    set,
    entries,
  }
}

const { saveDocumentInVectorStore, searchInVectorStore } = createVectorStoreTools(createPineconeVectorStore())

const exampleAgent = agent({
  role: '...',
  description: '...',
  tools: {
    saveDocumentInVectorStore,
    searchInVectorStore
  },
})

🖼️ Vision tool

packages/tools/src/vision.ts

Uses LLM as a OCR / Vision tool. Extract text or features from a picture.

Example

import { visionTool } from '@fabrice-ai/tools/vision'

const exampleAgent = agent({
  role: '...',
  description: '...',
  tools: {
    visionTool    
  },
})

🗄️ FileSystem tools

packages/tools/src/filesystem.ts

File system tools for reading, writing and listing files. The tools are sandboxed into workingDir for safety.

Example

import { createFileSystemTools } from '@fabrice-ai/tools/filesystem'

const workingDir = path.resolve(import.meta.dirname, '../assets/')

const { saveFile, readFile, listFilesFromDirectory } = createFileSystemTools({
  workingDir,
})

const exampleAgent = agent({
  role: '...',
  description: '...',
  tools: {
    visionTool    
  },
})