context-portal
If you are the rightful owner of context-portal and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
Context Portal MCP (ConPort) is a database-backed server designed to manage structured project context, enhancing AI assistants' ability to provide accurate and context-aware responses.
Context Portal (ConPort) serves as a project's memory bank, storing critical information such as decisions, tasks, and architectural patterns in a structured format. This tool is designed to help AI assistants and developer tools within IDEs understand specific software projects better. By building a project-specific knowledge base, ConPort allows AI to access and utilize relevant information quickly, improving the accuracy and helpfulness of responses. It replaces older file-based systems with a more reliable and queryable database backend, using SQLite for each workspace. ConPort supports multi-workspace environments and integrates tightly with IDEs through STDIO mode. It also includes vector embeddings for semantic search, enhancing Retrieval Augmented Generation (RAG) capabilities.
Features
- Structured context storage using SQLite, with one database per workspace.
- MCP server built with Python/FastAPI, supporting various IDEs and client interfaces.
- Dynamic project knowledge graph creation with explicit relationships between context items.
- Vector data storage and semantic search capabilities for advanced RAG.
- Prompt caching support for AI assistants with compatible LLM providers.
Tools
get_product_context
Retrieves the overall project goals, features, and architecture.
update_product_context
Updates the product context with full or partial content.
log_decision
Logs an architectural or implementation decision.
get_decisions
Retrieves logged decisions with optional filters.
search_decisions_fts
Full-text search across decision fields.