convez.ai
MCP Server
BrianLusina

image-research-assistant

BrianLusina/image-research-assistant

Ver no GitHub
InicialPluginScore · de 0 a 100

An Image Research AI powered Assistant

Instalação

$npx convez-ai BrianLusina/image-research-assistant

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
0
Forks
0
Acompanhando
0
Saúde
71%
Último commit
há 5 dias
Versão
—
Issues abertas
1
Contribuidores
2

README

A tradução para o português deste README ainda está na fila.

Image Research Assistant

This is an image research assistant focused on helping with image-related information retrieval and enrichment using modular services. The repository contains a lightweight application entry point, client/agent scaffolding, and service modules for vision and Wikipedia lookups.

Goals

  • Provide a compact, extensible architecture for image research tasks using MCP services.
  • Keep services modular so contributors can add new data sources or models.
  • Provide simple developer ergonomics for running and testing locally.

Contents

  • src/app/ — main application package and entry point
  • src/app/clients/ — client-side agent and MCP client code
  • src/app/servers/vision/ — vision-related server utilities and server
  • src/app/servers/wikipedia/ — wikipedia lookup server and helpers

Prerequisites

  • macOS/Linux/Windows with a POSIX-like shell (examples use zsh/bash)
  • Python 3.10 or newer
  • Git
  • (Optional) poetry if you prefer managing the project via Poetry. Though uv is the preferred package manager for this project.

Recommended setup

  1. Create and activate a virtual environment (macOS/Linux):
python3 -m venv .venv
source .venv/bin/activate

If using uv, simply run uv venv after installation to activate the environment.

  1. Install dependencies. Preferred project methods:
  • Use the Makefile helper (recommended):
make install
  • Or use uv (preferred package manager for this project) to install from pyproject.toml:
uv install -r pyproject.toml

If you prefer, you can still perform an editable pip install when iterating on local code:

pip install -e .

Note: There is no requirements.txt in this repository; use pyproject.toml with uv or the Makefile helper.

Running the project

  • Run the main application directly (quick):
python src/app/main.py
  • Or run as a module after editable install:
make run

Project considerations

  • The repository is intentionally modular: add new servers/ or clients/ components without touching core logic.
  • Keep third-party model or API keys out of the repo — use environment variables or a secrets store for credentials.
  • Aim for lightweight dependencies to keep the project easy to set up for contributors.
  • If adding long-running services, prefer running them behind a process manager (systemd, supervisor) or in containers for production.

Testing

  • There are no tests bundled by default. If you add tests, prefer pytest and place tests under tests/.
  • Example test run (after adding pytest to deps):
pytest -q

Contributing

  • Fork the repository and create feature branches: git checkout -b feat/your-feature
  • Keep commits small and focused, with clear messages.
  • Run linters and any tests you add before opening a PR.
  • Open a Pull Request against main, add a descriptive title and summary of changes.
  • For larger changes, open an issue first to discuss design and approach.

Code style

  • Follow idiomatic Python (PEP8). Optionally add black / ruff or similar tools to the project for consistency. Ruff is used here and can be run using make lint and type checking with make pyrefly.

Contact / Maintainers

  • The main repo owner is BrianLusina (GitHub: BrianLusina). Open issues or PRs for questions, fixes, or feature requests.
Ver o original no GitHub

Bom para

Relacionados