├── .gitignore
├── README.md
├── app.py
├── file_crew
    ├── .gitignore
    ├── README.md
    ├── knowledge
    │   └── user_preference.txt
    ├── pyproject.toml
    └── src
    │   └── file_crew
    │       ├── __init__.py
    │       ├── config
    │           ├── agents.yaml
    │           └── tasks.yaml
    │       ├── crew.py
    │       ├── main.py
    │       └── tools
    │           ├── __init__.py
    │           └── custom_tool.py
└── requirements.txt


/.gitignore:
--------------------------------------------------------------------------------
1 | __pycache__


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
  1 | # PDF Processing with Docling and CrewAI
  2 | 
  3 | This is a FastAPI application that uses Docling to parse PDF files and extract their content, then analyzes this content using CrewAI to generate structured reports and insights.
  4 | 
  5 | ## 5 Steps:
  6 | 1. Create Frontend with Bolt.new
  7 | 2. Create FastAPI Project
  8 | 3. Create Crew
  9 | 4. Ngrok Static Domain
 10 | 5. Connect Frontend with FastAPI
 11 | 
 12 | ## Features
 13 | 
 14 | - **Document Parsing**: Uses Docling to extract text content from PDF files
 15 | - **AI Analysis**: Processes extracted content with CrewAI to generate structured reports
 16 | - **API Interface**: Simple REST API for file upload and processing
 17 | - **Markdown Export**: Returns parsed content in markdown format
 18 | - **Clean Implementation**: Proper error handling and temporary file management
 19 | 
 20 | ## Requirements
 21 | 
 22 | - Python 3.9+
 23 | - FastAPI
 24 | - Docling
 25 | - CrewAI
 26 | - Other dependencies as specified in requirements.txt
 27 | 
 28 | ## Setup
 29 | 
 30 | 1. Clone the repository:
 31 | 
 32 | ```bash
 33 | git clone <your-repository-url>
 34 | cd <repository-directory>
 35 | ```
 36 | 
 37 | 2. Set up a virtual environment (recommended):
 38 | 
 39 | ```bash
 40 | python -m venv venv
 41 | source venv/bin/activate  # On Windows: venv\Scripts\activate
 42 | ```
 43 | 
 44 | 3. Install the required dependencies:
 45 | 
 46 | ```bash
 47 | pip install -r requirements.txt
 48 | ```
 49 | 
 50 | 4. Environment variables (optional):
 51 |    
 52 | If you want to customize the CrewAI behavior or use specific LLM models, you can create a `.env` file with appropriate settings:
 53 | 
 54 | ```
 55 | OPENAI_API_KEY=your_openai_api_key
 56 | MODEL_NAME=gpt-4-turbo
 57 | # Add other environment variables as needed
 58 | ```
 59 | 
 60 | 5. Run the application:
 61 | 
 62 | ```bash
 63 | python app.py
 64 | ```
 65 | 
 66 | The API will be available at `http://<your-local-ip>:5010`.
 67 | 
 68 | ## API Endpoints
 69 | 
 70 | ### Process PDF File
 71 | 
 72 | **Endpoint**: `/file-handler`  
 73 | **Method**: POST  
 74 | **Content-Type**: multipart/form-data
 75 | 
 76 | **Parameters**:
 77 | - `file`: The PDF file to parse and analyze
 78 | 
 79 | **Example using curl**:
 80 | ```bash
 81 | curl -X POST "http://localhost:5010/file-handler" \
 82 |   -H "accept: application/json" \
 83 |   -H "Content-Type: multipart/form-data" \
 84 |   -F "file=@/path/to/your/file.pdf"
 85 | ```
 86 | 
 87 | **Example using JavaScript/Fetch**:
 88 | ```javascript
 89 | const formData = new FormData();
 90 | formData.append('file', pdfFile); // pdfFile is a File object
 91 | 
 92 | fetch('http://localhost:5010/file-handler', {
 93 |   method: 'POST',
 94 |   body: formData
 95 | })
 96 | .then(response => response.json())
 97 | .then(data => console.log(data))
 98 | .catch(error => console.error('Error:', error));
 99 | ```
100 | 
101 | **Response**:
102 | ```json
103 | {
104 |   "filename": "example.pdf",
105 |   "markdown": "# Document Title\n\nContent in markdown format...",
106 |   "result": {
107 |     "key_points": ["Point 1", "Point 2", "..."],
108 |     "quick_summary": "Summary of the document...",
109 |     "extended_summary": "Detailed analysis...",
110 |     "actionable_insights": ["Insight 1", "Insight 2", "..."],
111 |     "source_documents": ["Document title and description..."],
112 |     "potential_biases": "Discussion of biases..."
113 |   }
114 | }
115 | ```
116 | 
117 | ## Implementation Details
118 | 
119 | ### How It Works
120 | 
121 | 1. **PDF Processing**:
122 |    - The application receives a PDF file through the `/file-handler` endpoint
123 |    - Docling's DocumentConverter processes the PDF and extracts structured content
124 |    - The content is converted to markdown format for further processing
125 | 
126 | 2. **AI Analysis**:
127 |    - The markdown content is passed to the CrewAI system
128 |    - A researcher agent analyzes the content according to the task definition
129 |    - The agent generates a structured report with key points, summaries, and insights
130 | 
131 | 3. **Result Delivery**:
132 |    - Both the original markdown and the AI-generated analysis are returned to the client
133 |    - The response includes detailed sections as specified in the task configuration
134 | 
135 | ### CrewAI Task Configuration
136 | 
137 | The CrewAI system is configured to analyze documents with a specific focus. The task configuration is defined in `file_crew/src/file_crew/config/tasks.yaml` and includes:
138 | 
139 | - Extracting key points from the document
140 | - Creating executive summaries
141 | - Providing detailed analysis
142 | - Suggesting actionable insights
143 | - Identifying potential biases in the source material
144 | 
145 | ## Testing
146 | 
147 | You can use the included test script to try out the API:
148 | 
149 | ```bash
150 | python test_pdf_parser.py /path/to/your/file.pdf
151 | ```
152 | 
153 | This will:
154 | 1. Upload the specified PDF to the API
155 | 2. Process the file through both Docling and CrewAI
156 | 3. Display a summary of the results in the terminal
157 | 4. Save the full results to a JSON file
158 | 
159 | ## Project Structure
160 | 
161 | ```
162 | .
163 | ├── app.py                  # FastAPI application
164 | ├── requirements.txt        # Project dependencies
165 | ├── test_pdf_parser.py      # Test script for the API
166 | ├── file_crew/
167 | │   └── src/
168 | │       └── file_crew/
169 | │           ├── main.py     # CrewAI integration
170 | │           └── config/
171 | │               └── tasks.yaml  # Task definitions for the AI analysis
172 | ```
173 | 
174 | ## Troubleshooting
175 | 
176 | - **PDF Parsing Issues**: Make sure the uploaded PDF is not corrupted and doesn't have DRM protection
177 | - **CrewAI Errors**: Check that any required environment variables are set correctly
178 | - **API Connection Issues**: Verify that the API is running and accessible from your client
179 | 
180 | 
181 | ## Acknowledgements
182 | 
183 | - [Docling](https://github.com/docling-project/docling) for PDF parsing
184 | - [CrewAI](https://github.com/joaomdmoura/crewAI) for AI agent orchestration
185 | 


--------------------------------------------------------------------------------
/app.py:
--------------------------------------------------------------------------------
  1 | from fastapi import FastAPI, HTTPException, Body, File, Form, UploadFile
  2 | from fastapi.middleware.cors import CORSMiddleware
  3 | from fastapi.responses import JSONResponse
  4 | import socket
  5 | import json
  6 | from typing import List, Dict, Any
  7 | import uvicorn
  8 | import tempfile
  9 | import os
 10 | from docling.document_converter import DocumentConverter
 11 | from file_crew.src.file_crew.main import run
 12 | 
 13 | def get_local_ip() -> str:
 14 |     """
 15 |     Get the local IP address of the machine.
 16 |     
 17 |     Returns:
 18 |         str: The local IP address, or '127.0.0.1' if it can't be determined.
 19 |     """
 20 |     try:
 21 |         s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
 22 |         s.connect(('8.8.8.8', 80))
 23 |         ip = s.getsockname()[0]
 24 |         s.close()
 25 |         return ip
 26 |     except Exception:
 27 |         return '127.0.0.1'
 28 | 
 29 | app = FastAPI(
 30 |     title="CrewAI API",
 31 |     description="API for processing documents with Docling and CrewAI",
 32 |     version="1.0.0"
 33 | )
 34 | 
 35 | # Configure CORS
 36 | app.add_middleware(
 37 |     CORSMiddleware,
 38 |     allow_origins=["*"],
 39 |     allow_credentials=True,
 40 |     allow_methods=["*"],
 41 |     allow_headers=["*"],
 42 | )
 43 | 
 44 | @app.post("/file-handler", response_model=Dict[str, Any])
 45 | async def file_handler_endpoint(
 46 |     file: UploadFile = File(...)
 47 | ):
 48 |     """
 49 |     Process a PDF file using Docling and CrewAI.
 50 |     
 51 |     Args:
 52 |         file (UploadFile): The PDF file to process
 53 |         
 54 |     Returns:
 55 |         Dict[str, Any]: The processed results including markdown content and CrewAI analysis
 56 |         
 57 |     Raises:
 58 |         HTTPException: If there's an error processing the file
 59 |     """
 60 |     temp_file_path = None
 61 |     
 62 |     try:
 63 |         print(f"Received file: {file.filename} (type: {file.content_type}, size: {file.size})")
 64 |         
 65 |         # Validate file type
 66 |         if not file.content_type.startswith('application/pdf'):
 67 |             raise HTTPException(
 68 |                 status_code=400, 
 69 |                 detail=f"Invalid file type: {file.content_type}. Please upload a PDF file."
 70 |             )
 71 |         
 72 |         # Create a temporary file to save the uploaded PDF
 73 |         with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as temp_file:
 74 |             # Write the uploaded file content to the temporary file
 75 |             content = await file.read()
 76 |             temp_file.write(content)
 77 |             temp_file_path = temp_file.name
 78 |         
 79 |         # Process the file
 80 |         try:
 81 |             # Initialize the DocumentConverter
 82 |             converter = DocumentConverter()
 83 |             
 84 |             # Convert the PDF document
 85 |             docling_result = converter.convert(temp_file_path)
 86 |             
 87 |             # Get the parsed content in markdown format
 88 |             markdown_content = docling_result.document.export_to_markdown()
 89 |             
 90 |             # Process with CrewAI
 91 |             crew_result = run(markdown_content)
 92 |             
 93 |             # Return the results
 94 |             return {
 95 |                 "filename": file.filename,
 96 |                 "markdown": markdown_content,
 97 |                 "result": crew_result.model_dump()
 98 |             }
 99 |         
100 |         finally:
101 |             # Always clean up the temporary file
102 |             if temp_file_path and os.path.exists(temp_file_path):
103 |                 os.unlink(temp_file_path)
104 |             
105 |     except HTTPException as e:
106 |         # Re-raise HTTP exceptions
107 |         raise e
108 |     except json.JSONDecodeError as e:
109 |         print(f"JSON decode error: {str(e)}")
110 |         raise HTTPException(status_code=400, detail=f"Invalid JSON payload: {str(e)}")
111 |     except Exception as e:
112 |         print(f"Error in file_handler: {str(e)}")
113 |         import traceback
114 |         traceback.print_exc()
115 |         raise HTTPException(status_code=500, detail=str(e))
116 | 
117 | @app.get("/")
118 | async def root():
119 |     """Root endpoint that returns basic API information."""
120 |     return {
121 |         "message": "Welcome to the CrewAI Document Processing API",
122 |         "endpoints": {
123 |             "/file-handler": "POST endpoint for processing PDF files"
124 |         }
125 |     }
126 | 
127 | if __name__ == "__main__":
128 |     host_ip = get_local_ip()
129 |     print(f"Starting server at http://{host_ip}:5010")
130 |     uvicorn.run("app:app", host=host_ip, port=5010, reload=True)
131 | 


--------------------------------------------------------------------------------
/file_crew/.gitignore:
--------------------------------------------------------------------------------
1 | .env
2 | __pycache__/
3 | .DS_Store
4 | 


--------------------------------------------------------------------------------
/file_crew/README.md:
--------------------------------------------------------------------------------
 1 | # FileCrew Crew
 2 | 
 3 | Welcome to the FileCrew Crew project, powered by [crewAI](https://crewai.com). This template is designed to help you set up a multi-agent AI system with ease, leveraging the powerful and flexible framework provided by crewAI. Our goal is to enable your agents to collaborate effectively on complex tasks, maximizing their collective intelligence and capabilities.
 4 | 
 5 | ## Installation
 6 | 
 7 | Ensure you have Python >=3.10 <3.13 installed on your system. This project uses [UV](https://docs.astral.sh/uv/) for dependency management and package handling, offering a seamless setup and execution experience.
 8 | 
 9 | First, if you haven't already, install uv:
10 | 
11 | ```bash
12 | pip install uv
13 | ```
14 | 
15 | Next, navigate to your project directory and install the dependencies:
16 | 
17 | (Optional) Lock the dependencies and install them by using the CLI command:
18 | ```bash
19 | crewai install
20 | ```
21 | ### Customizing
22 | 
23 | **Add your `OPENAI_API_KEY` into the `.env` file**
24 | 
25 | - Modify `src/file_crew/config/agents.yaml` to define your agents
26 | - Modify `src/file_crew/config/tasks.yaml` to define your tasks
27 | - Modify `src/file_crew/crew.py` to add your own logic, tools and specific args
28 | - Modify `src/file_crew/main.py` to add custom inputs for your agents and tasks
29 | 
30 | ## Running the Project
31 | 
32 | To kickstart your crew of AI agents and begin task execution, run this from the root folder of your project:
33 | 
34 | ```bash
35 | $ crewai run
36 | ```
37 | 
38 | This command initializes the file_crew Crew, assembling the agents and assigning them tasks as defined in your configuration.
39 | 
40 | This example, unmodified, will run the create a `report.md` file with the output of a research on LLMs in the root folder.
41 | 
42 | ## Understanding Your Crew
43 | 
44 | The file_crew Crew is composed of multiple AI agents, each with unique roles, goals, and tools. These agents collaborate on a series of tasks, defined in `config/tasks.yaml`, leveraging their collective skills to achieve complex objectives. The `config/agents.yaml` file outlines the capabilities and configurations of each agent in your crew.
45 | 
46 | ## Support
47 | 
48 | For support, questions, or feedback regarding the FileCrew Crew or crewAI.
49 | - Visit our [documentation](https://docs.crewai.com)
50 | - Reach out to us through our [GitHub repository](https://github.com/joaomdmoura/crewai)
51 | - [Join our Discord](https://discord.com/invite/X4JWnZnxPb)
52 | - [Chat with our docs](https://chatg.pt/DWjSBZn)
53 | 
54 | Let's create wonders together with the power and simplicity of crewAI.
55 | 


--------------------------------------------------------------------------------
/file_crew/knowledge/user_preference.txt:
--------------------------------------------------------------------------------
1 | User name is John Doe.
2 | User is an AI Engineer.
3 | User is interested in AI Agents.
4 | User is based in San Francisco, California.
5 | 


--------------------------------------------------------------------------------
/file_crew/pyproject.toml:
--------------------------------------------------------------------------------
 1 | [project]
 2 | name = "file_crew"
 3 | version = "0.1.0"
 4 | description = "file_crew using crewAI"
 5 | authors = [{ name = "Your Name", email = "you@example.com" }]
 6 | requires-python = ">=3.10,<3.13"
 7 | dependencies = [
 8 |     "crewai[tools]>=0.105.0,<1.0.0"
 9 | ]
10 | 
11 | [project.scripts]
12 | file_crew = "file_crew.main:run"
13 | run_crew = "file_crew.main:run"
14 | train = "file_crew.main:train"
15 | replay = "file_crew.main:replay"
16 | test = "file_crew.main:test"
17 | 
18 | [build-system]
19 | requires = ["hatchling"]
20 | build-backend = "hatchling.build"
21 | 
22 | [tool.crewai]
23 | type = "crew"
24 | 


--------------------------------------------------------------------------------
/file_crew/src/file_crew/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/tylerprogramming/crewai-frontend/b9035966a2a460c05104404122f7e03d0faf9ef5/file_crew/src/file_crew/__init__.py


--------------------------------------------------------------------------------
/file_crew/src/file_crew/config/agents.yaml:
--------------------------------------------------------------------------------
 1 | researcher:
 2 |   role: >
 3 |     Markdown Miner & Insight Extractor
 4 |   goal: >
 5 |      To efficiently and accurately extract critical information from markdown documents, identify key themes and relationships within the text, and provide the user with a concise and actionable summary of the document's contents. 
 6 |      The agent prioritizes accuracy and thoroughness, ensuring no vital detail is overlooked. {file_text}
 7 |   backstory: >
 8 |     Created to navigate the growing landscape of markdown-based knowledge, this agent is engineered for precision extraction and insightful synthesis. 
 9 |     Its core strength lies in its ability to deeply understand markdown structures, identify crucial data points, and distill them into actionable summaries. 
10 |     It acts as a dedicated research assistant, transforming raw markdown data into readily accessible and understandable information.
11 | 


--------------------------------------------------------------------------------
/file_crew/src/file_crew/config/tasks.yaml:
--------------------------------------------------------------------------------
 1 | research_task:
 2 |   description: |
 3 |     Parse and analyze markdown documents {file_text} to identify key developments and relevant information.
 4 |     Focus on extracting factual statements, key findings, and notable trends within the documents.
 5 |   expected_output: |
 6 |     A structured report containing the following sections, derived from the analyzed markdown documents:
 7 |     
 8 |     1. Key Points (Concise Summary):
 9 |        - A list of 5-7 bullet points summarizing the most critical and impactful findings related to {topic}.
10 |        - Each bullet point should be factually accurate and referenced to its source markdown document.
11 |        - Aim for clarity and conciseness, capturing the essence of the information in a single sentence.
12 |     
13 |     2. Quick Summary (Executive Overview):
14 |        - A brief paragraph (approximately 50-75 words) providing a high-level overview of the current state of {topic}.
15 |        - Highlight the most significant recent developments and their potential implications.
16 |        - Avoid jargon and technical terms, making it accessible to a general audience.
17 |     
18 |     3. Extended Summary (Detailed Analysis):
19 |        - A more in-depth summary (approximately 200-300 words) expanding on the Key Points and Quick Summary.
20 |        - Provide more context for the key findings, explaining their significance and potential impact.
21 |        - Include details on any conflicting viewpoints or ongoing debates within the analyzed documents.
22 |        - Maintain clear references to the source markdown documents for each key point.
23 |     
24 |     4. Actionable Insights (What We Should Take From This):
25 |        - A section outlining 3-5 specific actions or recommendations that can be derived from the research.
26 |        - These should be practical and based on the identified trends, findings, and potential implications.
27 |        - Examples: 
28 |          - "Further investigate the impact of {specific technology} on {specific application}"
29 |          - "Develop a strategy to address the emerging challenges related to {specific problem}"
30 |          - "Monitor the progress of {specific research area} due to its potential impact on {specific field}"
31 |        - For each actionable insight, explain the rationale behind the recommendation and its potential benefits.
32 |     
33 |     5. Source Document List:
34 |        - A comprehensive list of all markdown files analyzed, including their titles and brief descriptions.
35 |     
36 |     6. Potential Biases and Limitations:
37 |        - A brief paragraph acknowledging any potential biases or limitations inherent in the analyzed markdown data.
38 |        - This could include: 
39 |          - Overrepresentation of certain viewpoints
40 |          - Lack of coverage in specific areas
41 |          - Potential for outdated information
42 |          - Acknowledge if the Markdown files come from a specific source and therefore might have a bias
43 |   agent: researcher
44 | 
45 | 


--------------------------------------------------------------------------------
/file_crew/src/file_crew/crew.py:
--------------------------------------------------------------------------------
 1 | from crewai import Agent, Crew, Process, Task, LLM
 2 | from crewai.project import CrewBase, agent, crew, task
 3 | from dotenv import load_dotenv
 4 | import os
 5 | from pydantic import BaseModel
 6 | from typing import List
 7 | 
 8 | load_dotenv()
 9 | 
10 | google_flash_llm = LLM(
11 | 	model="gemini/gemini-2.0-flash",
12 | 	api_key=os.getenv("GOOGLE_API_KEY"),
13 | )
14 | 
15 | class ResearchOutput(BaseModel):
16 |     summary: str
17 |     key_points: List[str]
18 |     quick_summary: str
19 |     extended_summary: str
20 |     actionable_insights: List[str]
21 |     source_document_list: List[str]
22 |     potential_biases_and_limitations: str
23 | 
24 | @CrewBase
25 | class FileCrew():
26 | 
27 | 	agents_config = 'config/agents.yaml'
28 | 	tasks_config = 'config/tasks.yaml'
29 | 
30 | 	@agent
31 | 	def researcher(self) -> Agent:
32 | 		return Agent(
33 | 			config=self.agents_config['researcher'],
34 | 			verbose=True,
35 | 			llm=google_flash_llm
36 | 		)
37 | 
38 | 	@task
39 | 	def research_task(self) -> Task:
40 | 		return Task(
41 | 			config=self.tasks_config['research_task'],
42 | 			output_pydantic=ResearchOutput
43 | 		)
44 | 
45 | 	@crew
46 | 	def crew(self) -> Crew:
47 | 		"""Creates the FileCrew crew"""
48 | 
49 | 		return Crew(
50 | 			agents=self.agents,
51 | 			tasks=self.tasks,
52 | 			process=Process.sequential,
53 | 			verbose=True,
54 | 		)
55 | 


--------------------------------------------------------------------------------
/file_crew/src/file_crew/main.py:
--------------------------------------------------------------------------------
 1 | #!/usr/bin/env python
 2 | import sys
 3 | import warnings
 4 | 
 5 | from datetime import datetime
 6 | 
 7 | from .crew import FileCrew
 8 | 
 9 | warnings.filterwarnings("ignore", category=SyntaxWarning, module="pysbd")
10 | 
11 | # This main file is intended to be a way for you to run your
12 | # crew locally, so refrain from adding unnecessary logic into this file.
13 | # Replace with inputs you want to test with, it will automatically
14 | # interpolate any tasks and agents information
15 | 
16 | def run(file_text):
17 |     """
18 |     Run the crew.
19 |     """
20 |     inputs = {
21 |         'file_text': file_text,
22 |     }
23 |     
24 |     try:
25 |         result = FileCrew().crew().kickoff(inputs=inputs)
26 |         return result.pydantic
27 |     except Exception as e:
28 |         raise Exception(f"An error occurred while running the crew: {e}")
29 | 
30 | 
31 | def train():
32 |     """
33 |     Train the crew for a given number of iterations.
34 |     """
35 |     inputs = {
36 |         "topic": "AI LLMs"
37 |     }
38 |     try:
39 |         FileCrew().crew().train(n_iterations=int(sys.argv[1]), filename=sys.argv[2], inputs=inputs)
40 | 
41 |     except Exception as e:
42 |         raise Exception(f"An error occurred while training the crew: {e}")
43 | 
44 | def replay():
45 |     """
46 |     Replay the crew execution from a specific task.
47 |     """
48 |     try:
49 |         FileCrew().crew().replay(task_id=sys.argv[1])
50 | 
51 |     except Exception as e:
52 |         raise Exception(f"An error occurred while replaying the crew: {e}")
53 | 
54 | def test():
55 |     """
56 |     Test the crew execution and returns the results.
57 |     """
58 |     inputs = {
59 |         "topic": "AI LLMs",
60 |         "current_year": str(datetime.now().year)
61 |     }
62 |     try:
63 |         FileCrew().crew().test(n_iterations=int(sys.argv[1]), openai_model_name=sys.argv[2], inputs=inputs)
64 | 
65 |     except Exception as e:
66 |         raise Exception(f"An error occurred while testing the crew: {e}")
67 | 


--------------------------------------------------------------------------------
/file_crew/src/file_crew/tools/__init__.py:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/tylerprogramming/crewai-frontend/b9035966a2a460c05104404122f7e03d0faf9ef5/file_crew/src/file_crew/tools/__init__.py


--------------------------------------------------------------------------------
/file_crew/src/file_crew/tools/custom_tool.py:
--------------------------------------------------------------------------------
 1 | from crewai.tools import BaseTool
 2 | from typing import Type
 3 | from pydantic import BaseModel, Field
 4 | 
 5 | 
 6 | class MyCustomToolInput(BaseModel):
 7 |     """Input schema for MyCustomTool."""
 8 |     argument: str = Field(..., description="Description of the argument.")
 9 | 
10 | class MyCustomTool(BaseTool):
11 |     name: str = "Name of my tool"
12 |     description: str = (
13 |         "Clear description for what this tool is useful for, your agent will need this information to use it."
14 |     )
15 |     args_schema: Type[BaseModel] = MyCustomToolInput
16 | 
17 |     def _run(self, argument: str) -> str:
18 |         # Implementation goes here
19 |         return "this is an example of a tool output, ignore it and move along."
20 | 


--------------------------------------------------------------------------------
/requirements.txt:
--------------------------------------------------------------------------------
  1 | agentops==0.3.26
  2 | agentstack==0.3.5
  3 | aiofiles==24.1.0
  4 | aiohappyeyeballs==2.4.4
  5 | aiohttp==3.11.10
  6 | aiohttp-retry==2.8.3
  7 | aiosignal==1.3.2
  8 | aiosqlite==0.21.0
  9 | alembic==1.14.0
 10 | altair==5.5.0
 11 | annotated-types==0.7.0
 12 | anyio==4.7.0
 13 | appdirs==1.4.4
 14 | arrow==1.3.0
 15 | art==6.4
 16 | asgiref==3.8.1
 17 | asn1crypto==1.5.1
 18 | astor==0.8.1
 19 | asttokens==3.0.0
 20 | atpublic==5.0
 21 | attrs==24.2.0
 22 | auth0-python==4.7.2
 23 | Authlib==1.3.1
 24 | autoflake==2.3.1
 25 | babel==2.17.0
 26 | backoff==2.2.1
 27 | backrefs==5.8
 28 | bcrypt==4.2.1
 29 | beautifulsoup4==4.12.3
 30 | binaryornot==0.4.4
 31 | blessed==1.20.0
 32 | blinker==1.9.0
 33 | build==1.2.2.post1
 34 | cached-property==2.0.1
 35 | cachetools==5.5.0
 36 | certifi==2024.12.14
 37 | cffi==1.17.1
 38 | chardet==5.2.0
 39 | charset-normalizer==3.4.0
 40 | chroma-hnswlib==0.7.6
 41 | chromadb==0.5.23
 42 | click==8.1.8
 43 | cohere==5.13.3
 44 | colorama==0.4.6
 45 | coloredlogs==15.0.1
 46 | composio==0.1.1
 47 | composio_core==0.7.2
 48 | composio_crewai==0.7.2
 49 | composio_langchain==0.7.2
 50 | cookiecutter==2.6.0
 51 | Crawl4AI==0.4.248
 52 | crewai==0.105.0
 53 | crewai-tools==0.37.0
 54 | cryptography==43.0.3
 55 | css-inline==0.14.3
 56 | cssselect==1.2.0
 57 | dataclasses-json==0.6.7
 58 | dateparser==1.2.0
 59 | decorator==5.1.1
 60 | deepsearch-glm==1.0.0
 61 | Deprecated==1.2.15
 62 | deprecation==2.1.0
 63 | dill==0.3.9
 64 | dirtyjson==1.0.8
 65 | distro==1.9.0
 66 | docker==7.1.0
 67 | docling==2.15.1
 68 | docling-core==2.14.0
 69 | docling-ibm-models==3.1.2
 70 | docling-parse==3.0.0
 71 | docstring_parser==0.16
 72 | docx2txt==0.8
 73 | durationpy==0.9
 74 | easyocr==1.7.2
 75 | editor==1.6.6
 76 | elevenlabs==1.53.0
 77 | embedchain==0.1.125
 78 | et_xmlfile==2.0.0
 79 | exa-py==1.8.8
 80 | executing==2.1.0
 81 | fake-http-header==0.3.5
 82 | fake-useragent==2.0.3
 83 | fastapi==0.115.11
 84 | fastavro==1.9.7
 85 | fastembed==0.5.1
 86 | feedparser==6.0.11
 87 | ffmpeg==1.4
 88 | ffmpy==0.5.0
 89 | filelock==3.16.1
 90 | filetype==1.2.0
 91 | firecrawl-py==1.6.4
 92 | Flask==3.1.0
 93 | Flask-Cors==5.0.0
 94 | flatbuffers==24.3.25
 95 | frozenlist==1.5.0
 96 | fsspec==2024.10.0
 97 | ghp-import==2.1.0
 98 | gitdb==4.0.12
 99 | GitPython==3.1.44
100 | google-ai-generativelanguage==0.6.15
101 | google-api-core==2.24.0
102 | google-api-python-client==2.154.0
103 | google-auth==2.37.0
104 | google-auth-httplib2==0.2.0
105 | google-auth-oauthlib==1.2.1
106 | google-cloud-aiplatform==1.74.0
107 | google-cloud-bigquery==3.27.0
108 | google-cloud-core==2.4.1
109 | google-cloud-resource-manager==1.14.0
110 | google-cloud-storage==2.19.0
111 | google-crc32c==1.6.0
112 | google-genai==1.0.0
113 | google-generativeai==0.8.4
114 | google-resumable-media==2.7.2
115 | google_news_feed==1.1.0
116 | googleapis-common-protos==1.66.0
117 | gotrue==2.11.0
118 | gptcache==0.1.44
119 | gradio==5.13.1
120 | gradio_client==1.6.0
121 | greenlet==3.1.1
122 | griffe==1.6.0
123 | grpc-google-iam-v1==0.13.1
124 | grpcio==1.69.0
125 | grpcio-health-checking==1.69.0
126 | grpcio-status==1.68.1
127 | grpcio-tools==1.68.1
128 | h11==0.14.0
129 | h2==4.1.0
130 | hpack==4.0.0
131 | httpcore==1.0.7
132 | httplib2==0.22.0
133 | httptools==0.6.4
134 | httpx==0.27.2
135 | httpx-sse==0.4.0
136 | huggingface-hub==0.26.5
137 | humanfriendly==10.0
138 | hyperframe==6.0.1
139 | idna==3.10
140 | ijson==3.3.0
141 | imageio==2.36.1
142 | imageio-ffmpeg==0.6.0
143 | importlib_metadata==8.5.0
144 | importlib_resources==6.4.5
145 | inflection==0.5.1
146 | iniconfig==2.0.0
147 | inquirer==3.4.0
148 | instructor==1.7.0
149 | ipython==8.30.0
150 | itsdangerous==2.2.0
151 | jedi==0.19.2
152 | Jinja2==3.1.4
153 | jiter==0.6.1
154 | joblib==1.4.2
155 | json5==0.10.0
156 | json_repair==0.31.0
157 | jsonlines==3.1.0
158 | jsonpatch==1.33
159 | jsonpickle==4.0.1
160 | jsonpointer==3.0.0
161 | jsonref==1.1.0
162 | jsonschema==4.23.0
163 | jsonschema-specifications==2024.10.1
164 | kubernetes==31.0.0
165 | lancedb==0.17.0
166 | langchain==0.3.12
167 | langchain-cohere==0.3.3
168 | langchain-community==0.3.12
169 | langchain-core==0.3.25
170 | langchain-experimental==0.3.3
171 | langchain-openai==0.2.12
172 | langchain-text-splitters==0.3.3
173 | langchainhub==0.1.21
174 | langsmith==0.1.147
175 | lazy_loader==0.4
176 | linkup-sdk==0.2.1
177 | litellm==1.60.2
178 | llama-cloud==0.1.12
179 | llama-cloud-services==0.6.1
180 | llama-index-core==0.12.17
181 | llama-index-readers-file==0.4.5
182 | llvmlite==0.43.0
183 | loguru==0.7.3
184 | lxml==5.3.0
185 | Mako==1.3.8
186 | Markdown==3.7
187 | markdown-it-py==3.0.0
188 | marko==2.1.2
189 | MarkupSafe==2.1.5
190 | marshmallow==3.23.1
191 | matplotlib-inline==0.1.7
192 | mdurl==0.1.2
193 | mem0ai==0.1.35
194 | mergedeep==1.3.4
195 | mkdocs==1.6.1
196 | mkdocs-autorefs==1.4.1
197 | mkdocs-get-deps==0.2.0
198 | mkdocs-material==9.6.7
199 | mkdocs-material-extensions==1.3.1
200 | mkdocstrings==0.28.3
201 | mkdocstrings-python==1.16.3
202 | mmh3==4.1.0
203 | monotonic==1.6
204 | more-itertools==10.5.0
205 | moviepy==2.1.2
206 | mpire==2.10.2
207 | mpmath==1.3.0
208 | multidict==6.1.0
209 | multiprocess==0.70.17
210 | mypy-extensions==1.0.0
211 | narwhals==1.22.0
212 | nest-asyncio==1.6.0
213 | networkx==3.4.2
214 | ninja==1.11.1.3
215 | nltk==3.9.1
216 | nodeenv==1.9.1
217 | notion-client==2.3.0
218 | numba==0.60.0
219 | numpy==1.26.4
220 | oauthlib==3.2.2
221 | onnxruntime==1.20.1
222 | openai==1.66.2
223 | openai-agents==0.0.3
224 | openai-whisper==20240930
225 | opencv-python-headless==4.10.0.84
226 | openpyxl==3.1.5
227 | opentelemetry-api==1.29.0
228 | opentelemetry-exporter-otlp-proto-common==1.29.0
229 | opentelemetry-exporter-otlp-proto-grpc==1.29.0
230 | opentelemetry-exporter-otlp-proto-http==1.29.0
231 | opentelemetry-instrumentation==0.50b0
232 | opentelemetry-instrumentation-asgi==0.50b0
233 | opentelemetry-instrumentation-fastapi==0.50b0
234 | opentelemetry-proto==1.29.0
235 | opentelemetry-sdk==1.29.0
236 | opentelemetry-semantic-conventions==0.50b0
237 | opentelemetry-util-http==0.50b0
238 | orjson==3.10.12
239 | outcome==1.3.0.post0
240 | overrides==7.7.0
241 | packaging==23.2
242 | paginate==0.5.7
243 | pandas==2.2.3
244 | parameterized==0.9.0
245 | paramiko==3.5.1
246 | parso==0.8.4
247 | pathspec==0.12.1
248 | patronus==0.0.17
249 | pdfminer.six==20231228
250 | pdfplumber==0.11.4
251 | pexpect==4.9.0
252 | pillow==10.4.0
253 | platformdirs==4.3.6
254 | playwright==1.49.1
255 | plotly==6.0.0
256 | pluggy==1.5.0
257 | portalocker==2.10.1
258 | portkey-ai==1.10.2
259 | postgrest==0.18.0
260 | posthog==3.7.4
261 | proglog==0.1.10
262 | prompt_toolkit==3.0.48
263 | propcache==0.2.1
264 | proto-plus==1.25.0
265 | protobuf==5.29.1
266 | psutil==5.9.8
267 | ptyprocess==0.7.0
268 | pure_eval==0.2.3
269 | py_rust_stemmers==0.1.5
270 | pyairtable==3.0.0
271 | pyarrow==18.1.0
272 | pyasn1==0.6.1
273 | pyasn1_modules==0.4.1
274 | pyclipper==1.3.0.post6
275 | pycparser==2.22
276 | pydantic==2.10.3
277 | pydantic-settings==2.7.0
278 | pydantic_core==2.27.1
279 | pydeck==0.9.1
280 | pydub==0.25.1
281 | pyee==12.0.0
282 | pyflakes==3.2.0
283 | pygame==2.6.1
284 | Pygments==2.18.0
285 | PyJWT==2.10.1
286 | pylance==0.20.0
287 | pymdown-extensions==10.14.3
288 | PyNaCl==1.5.0
289 | pyOpenSSL==24.3.0
290 | pyparsing==3.2.0
291 | pypdf==5.1.0
292 | pypdfium2==4.30.0
293 | pyperclip==1.9.0
294 | PyPika==0.48.9
295 | pyproject_hooks==1.2.0
296 | pyright==1.1.390
297 | pysbd==0.3.4
298 | Pysher==1.0.8
299 | PySocks==1.7.1
300 | pytest==8.3.4
301 | python-bidi==0.6.3
302 | python-dateutil==2.9.0.post0
303 | python-docx==1.1.2
304 | python-dotenv==1.0.1
305 | python-multipart==0.0.20
306 | python-pptx==1.0.2
307 | python-slugify==8.0.4
308 | python-telegram-bot==21.10
309 | pytube==15.0.0
310 | pytz==2024.2
311 | pyvis==0.3.2
312 | PyYAML==6.0.2
313 | pyyaml_env_tag==0.1
314 | qdrant-client==1.12.1
315 | rank-bm25==0.2.2
316 | readchar==4.2.1
317 | realtime==2.0.6
318 | referencing==0.35.1
319 | regex==2024.11.6
320 | requests==2.32.3
321 | requests-oauthlib==2.0.0
322 | requests-toolbelt==1.0.0
323 | rich==13.9.4
324 | rpds-py==0.22.3
325 | rsa==4.9
326 | Rtree==1.3.0
327 | ruamel.yaml==0.18.6
328 | ruamel.yaml.base==0.3.2
329 | ruamel.yaml.clib==0.2.12
330 | ruff==0.9.3
331 | runs==1.2.2
332 | safehttpx==0.1.6
333 | safetensors==0.5.2
334 | schedule==1.2.2
335 | schema==0.7.7
336 | scikit-image==0.25.0
337 | scipy==1.15.1
338 | scrapegraph_py==1.8.0
339 | selenium==4.27.1
340 | semantic-version==2.10.0
341 | semchunk==2.2.2
342 | semver==3.0.4
343 | sentry-sdk==2.21.0
344 | serpapi==0.1.5
345 | sgmllib3k==1.0.0
346 | shapely==2.0.6
347 | shellingham==1.5.4
348 | six==1.17.0
349 | smmap==5.0.2
350 | sniffio==1.3.1
351 | snowballstemmer==2.2.0
352 | snowflake==1.0.2
353 | snowflake-connector-python==3.12.4
354 | snowflake._legacy==1.0.0
355 | snowflake.core==1.0.2
356 | sortedcontainers==2.4.0
357 | soupsieve==2.6
358 | spider-client==0.1.25
359 | SQLAlchemy==2.0.36
360 | stack-data==0.6.3
361 | starlette==0.41.3
362 | storage3==0.9.0
363 | streamlit==1.41.1
364 | striprtf==0.0.26
365 | supabase==2.10.0
366 | supafunc==0.7.0
367 | sympy==1.13.3
368 | tabulate==0.9.0
369 | tenacity==9.0.0
370 | termcolor==2.4.0
371 | text-unidecode==1.3
372 | tf-playwright-stealth==1.1.1
373 | tifffile==2025.1.10
374 | tiktoken==0.7.0
375 | tokenizers==0.19.1
376 | toml==0.10.2
377 | tomli==2.2.1
378 | tomli_w==1.1.0
379 | tomlkit==0.13.2
380 | torch==2.2.2
381 | torchvision==0.17.2
382 | tornado==6.4.2
383 | tqdm==4.67.1
384 | traitlets==5.14.3
385 | transformers==4.42.4
386 | trio==0.27.0
387 | trio-websocket==0.11.1
388 | twilio==9.4.3
389 | typer==0.12.5
390 | types-python-dateutil==2.9.0.20241206
391 | types-requests==2.32.0.20241016
392 | typing-inspect==0.9.0
393 | typing_extensions==4.12.2
394 | tzdata==2024.2
395 | tzlocal==5.2
396 | uritemplate==4.1.1
397 | urllib3==2.2.3
398 | uv==0.5.9
399 | uvicorn==0.34.0
400 | uvloop==0.21.0
401 | validators==0.34.0
402 | watchdog==6.0.0
403 | watchfiles==1.0.3
404 | wcwidth==0.2.13
405 | weaviate-client==4.10.2
406 | websocket-client==1.8.0
407 | websockets==13.1
408 | Werkzeug==3.1.3
409 | whisper==1.1.10
410 | wrapt==1.17.0
411 | wsproto==1.2.0
412 | XlsxWriter==3.2.0
413 | xmod==1.8.1
414 | xxhash==3.5.0
415 | yarl==1.18.3
416 | yt-dlp==2024.11.18
417 | zipp==3.21.0
418 | 


--------------------------------------------------------------------------------