Integuru: what it is, what problem it solves & why it's gaining traction

What it solves

Integuru v0 allows users to create automated integrations for platforms that do not have official APIs. It does this by reverse-engineering the network requests a browser makes when a user performs a specific action manually.

How it works

The tool uses a combination of browser network logs (HAR files) and LLMs to automate a process:

1. Capture: The user records their browser activity using a script to generate a HAR file and a cookie file.
2. Analysis: The agent analyzes the network requests to identify the target action (e.g., downloading a bill) and maps out the dependencies. If a request requires a dynamic ID (like a user ID) from a previous request, the agent identifies that source.
3. Graph Generation: It builds a dependency graph of all necessary requests, starting from the final action and working backward to the initial authentication step.
4. Code Generation: The agent converts this graph into runnable Python code that can execute the sequence of requests using the captured cookies.

Who it’s for

Developers and automation engineers who need to programmatically interact with websites or services that lack public API documentation.

Highlights

• Automated API Reverse-Engineering: Turns browser network traffic into functional code.
• Dependency Mapping: Automatically identifies and links dynamic parameters between requests.
• LLM-Powered: Utilizes high-reasoning models like GPT-4o and o1-preview for graph and code generation.
• Session Handling: Supports the use of cookies and authentication tokens to bypass manual login steps, including those following 2FA.