Perplexica MCP Server

# Roocode Rules for Perplexica MCP Server ## Project Overview This is a Model Context Protocol (MCP) server that provides AI assistants with access to Perplexica's AI-powered search engine. The server acts as a proxy between MCP clients and the Perplexica API. ## General - Maintain a task.md file for high level tasks that need step by step implementation - Use context7 MCP tool to always gather latest documentation / knowledge about a library - Use searxng tool to search web pages when necessary - Use mcp-server-firecrawl tool to scrape web pages when necessary - Use github tool for actions on remote repositories - Use git tool for actions on local repositories ## Testing - Write unit tests for all core functionality - Test MCP tool registration and execution - Mock external API calls in tests - Maintain test coverage above 80% ## Miscellaneous - Keep README.md up-to-date with installation and usage instructions - Document all environment variables in README.md - Document all configuration options in README.md - Update CHANGELOG.md with release notes - Use conventional commit format: `type(scope): description (issue#)` - Use .env files for local development - Follow semantic versioning (SemVer) - Update version in pyproject.toml - Tag releases in Git - Use `uv` for dependency management - Always use `uv run` for running Python implementations