Composio AI Product Ops Assignment

Submitted by: Garima Varma

Overview

This project implements an autonomous research agent using the Composio SDK to analyze developer platforms and determine whether they can be integrated as AI agent toolkits.

For each application, the agent:

1. Searches official developer documentation using Composio Search.
2. Extracts structured metadata using Groq GPT-OSS-120B.
3. Stores evidence-backed JSON results.
4. Performs verification of extracted information.
5. Generates an interactive HTML dashboard summarizing the findings.

Tech Stack

• Python
• Composio SDK
• Groq API (GPT-OSS-120B)
• HTML / CSS / JavaScript
• Chart.js

Project Structure

apps/
docs/
output/
src/
verification/
requirements.txt
README.md

How to Run

Install dependencies:

pip install -r requirements.txt

Run the research pipeline:

python src/main.py

Generate the final dashboard:

python src/generate_report.py

Open the generated report:

docs/index.html

using VS Code Live Server or any local web server.

Output

The project generates:

• Individual JSON results for each researched application
• Summary statistics and pattern analysis
• Interactive HTML dashboard
• Verification artifacts

Methodology

The research pipeline follows these steps:

1. Search official developer documentation using Composio Search.
2. Retrieve relevant documentation pages.
3. Extract structured metadata using Groq GPT-OSS-120B.
4. Save evidence-backed JSON outputs.
5. Aggregate all results.
6. Generate an interactive HTML report.

Limitations

• Some applications have limited or partner-only documentation.
• API availability may change over time.
• MCP availability is determined using publicly available documentation.
• A small number of applications may require manual verification where documentation is incomplete.

Repository

https://github.com/Garima040106/ComposioPI