# What Are the Backend Services for the i-have-adhd Repository? > Discover the backend services for the i-have-adhd repository. Learn how it leverages OpenAI and Google Gemini for processing without custom server logic. Understand its architecture now. - Repository: [Ayoub Ghriss/i-have-adhd](https://github.com/ayghri/i-have-adhd) - Tags: how-to-guide - Published: 2026-07-30 --- **The i-have-adhd project delegates all processing to external LLM providers—specifically OpenAI and Google Gemini—and contains no custom server-side logic, databases, or internal REST APIs.** The i-have-adhd repository functions as a skill plugin for AI coding assistants rather than hosting its own **backend services**. According to the ayghri/i-have-adhd source code analysis, the project delegates all processing to external large language model providers through dedicated configuration files. ## Architecture Overview: Plugin-Based Backend Delegation Unlike traditional applications that host backend services, i-have-adhd operates as a **skill/plugin** within compatible agent environments. The repository contains no server-side application code, database schemas, or message queues. Instead, the runtime environment—whether Claude Code, Codex, or another compatible agent—loads the plugin via [`plugin.json`](https://github.com/ayghri/i-have-adhd/blob/main/plugin.json) and routes requests to the configured LLM provider. The skill forwards all processing to external APIs defined in agent configuration files located in the `skills/i-have-adhd/agents/` directory. ## Primary Backend Services The repository supports two distinct LLM backend services, selectable through configuration files. ### OpenAI (GPT-4 / ChatGPT Models) The default backend service connects to **OpenAI's API** for generative AI processing. Configuration resides in: ```yaml # skills/i-have-adhd/agents/openai.yaml ``` This YAML file contains connection parameters including model name, temperature settings, and other inference controls required to generate ADHD-friendly responses through OpenAI's GPT models. ### Google Gemini As an alternative backend, the skill supports **Google Gemini** for users who prefer Google's LLM infrastructure. Configuration resides in: ```toml # skills/i-have-adhd/agents/gemini.toml ``` The TOML file provides equivalent connection settings for the Gemini API, allowing seamless swapping between OpenAI and Google backends without modifying the core skill logic. ## Key Configuration Files and Their Roles The repository delegates backend functionality through these specific files: - **[`skills/i-have-adhd/agents/openai.yaml`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/agents/openai.yaml)** — Defines OpenAI API connection parameters and model configuration - **[`skills/i-have-adhd/agents/gemini.toml`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/agents/gemini.toml)** — Defines Google Gemini API connection parameters - **[`plugin.json`](https://github.com/ayghri/i-have-adhd/blob/main/plugin.json)** — Metadata enabling Claude Code and Codex to load and execute the skill - **[`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md)** — Behavioral definitions and rules governing how the skill interacts with backend services No additional backend infrastructure exists in the repository beyond these configuration definitions. ## Runtime Integration: Claude Code and Codex The skill integrates with coding assistants through standardized commands. When invoked, the runtime environment reads the appropriate configuration file and routes the request to the selected backend service. **Using with Claude Code:** ```text /i-have-adhd ``` **Using with Codex:** ```text $i-have-adhd ``` Both commands trigger the same underlying logic—the runtime loads the skill, reads either [`openai.yaml`](https://github.com/ayghri/i-have-adhd/blob/main/openai.yaml) (default) or [`gemini.toml`](https://github.com/ayghri/i-have-adhd/blob/main/gemini.toml) (if switched), and forwards the request to the respective LLM provider's API. ## Summary - The i-have-adhd repository functions as a **skill plugin**, not a standalone application with hosted backend services. - **Two external LLM providers** serve as the actual backends: OpenAI and Google Gemini. - Configuration occurs through **[`skills/i-have-adhd/agents/openai.yaml`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/agents/openai.yaml)** and **[`skills/i-have-adhd/agents/gemini.toml`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/agents/gemini.toml)**. - The runtime environment (Claude Code, Codex, or compatible agents) handles API routing; the repository itself contains only configuration and behavioral definitions. - No databases, message queues, or custom REST services exist within the codebase. ## Frequently Asked Questions ### Does i-have-adhd host its own backend server? No. The repository contains no server-side application code, HTTP endpoints, or listening services. It operates entirely as a client-side skill that delegates processing to external LLM APIs through configuration files. ### How do I switch between OpenAI and Gemini backends? Switch backends by configuring your runtime environment to use the desired configuration file. The skill reads from [`skills/i-have-adhd/agents/openai.yaml`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/agents/openai.yaml) by default, but you can direct the runtime to use [`skills/i-have-adhd/agents/gemini.toml`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/agents/gemini.toml) instead to route requests to Google's Gemini API. ### What data is stored in the backend configuration files? The YAML and TOML files contain API connection parameters including model names (e.g., GPT-4 or Gemini variants), temperature settings, and authentication details required to communicate with the respective LLM providers. No user data or conversation history persists in these files. ### Can I add additional LLM providers as backends? While the repository currently defines configurations only for OpenAI and Gemini, the plugin architecture supports extension. You would need to create a new configuration file following the established patterns in `skills/i-have-adhd/agents/` and update the runtime environment to recognize the new provider endpoint.