Empower your development workflow with intelligent context awareness - DevContext understands your codebase, conversations, and development patterns to provide relevant context exactly when you need it.
DevContext is a cutting-edge Model Context Protocol (MCP) server designed to provide developers with continuous, project-centric context awareness. Unlike traditional context systems, DevContext leverages sophisticated retrieval methods, focusing on keyword analysis, relationship graphs, and structured metadata to deliver highly relevant context during development.
The server operates with a database instance dedicated to a single project, eliminating cross-project complexity and ensuring performance with minimal resource requirements. DevContext builds a comprehensive understanding of your codebase - from repository structure down to individual functions - while continuously learning from and adapting to your development patterns.
The BEST way to use this MCP server is to follow the guide below on implementing the provided Cursor Rules system which in turn gives you:
- Completely autonomous context management
- Autonomous external documentation context and use
- Complete task management workflow integration
- Node.js: Runtime environment (Node.js 18+)
- TursoDB: SQL database optimized for edge deployment (compatible with SQLite)
- Model Context Protocol SDK: For standardized communication with IDE clients
- Cursor Rules: Autonomous development environment and workflow management
- JavaScript/TypeScript: Pure JavaScript implementation with no external ML dependencies
- Node.js 18.0.0 or higher
- Cursor IDE with MCP support
- TursoDB account (for database)
-
Sign up for TursoDB:
- Visit Turso and create an account
- The free tier is sufficient for most projects
-
Install Turso CLI (optional but recommended):
curl -sSfL https://get.turso.tech/install.sh | bash -
Authenticate with Turso:
turso auth login
-
Create a project database:
turso db create devcontext
-
Get database credentials:
# Get database URL turso db show devcontext --url # Create auth token turso db tokens create devcontext
Save both the URL and token for the next step.
Create or edit .cursor/mcp.json in your project directory:
{
"mcpServers": {
"devcontext": {
"command": "npx",
"args": ["-y", "devcontext@latest"],
"enabled": true,
"env": {
"TURSO_DATABASE_URL": "your-turso-database-url",
"TURSO_AUTH_TOKEN": "your-turso-auth-token"
}
}
}
}Replace your-turso-database-url and your-turso-auth-token with the values obtained in Step 1.
DevContext implements a sophisticated set of Cursor Rules that create an autonomous development environment. These rules guide Cursor's AI assistants in maintaining project scope alignment, incorporating up-to-date documentation, and implementing advanced task workflows.
Be on the lookout for the DevContext Project Generator which is coming very soon and will create a COMPLETE set up for your project to literally 10x your development workflow.
The core rule defines a precise sequence for tool execution:
1. FIRST: Call initialize_conversation_context EXACTLY ONCE at START
2. AS NEEDED: Call update_conversation_context for code changes/new messages
3. AS NEEDED: Call retrieve_relevant_context when specific context is required
4. OCCASIONALLY: Call record_milestone_context for significant achievements
5. LAST: Call finalize_conversation_context EXACTLY ONCE at END
This workflow ensures comprehensive context management throughout the entire development session.
All external library usage must be preceded by proper documentation retrieval:
- Two-Step Documentation Retrieval using Context7
- Web Search Fallback for documentation not available through Context7
- Multi-Source Documentation Synthesis for comprehensive understanding
This prevents common issues with incorrect API usage, incompatible versions, or missing dependencies.
The task workflow system enables:
- Structured task management in
tasks.md - Task ID-based implementation order
- Status tracking with completion metadata
- Project blueprint integration for architectural context
-
Create Rules Directory:
mkdir -p .cursor/rules
-
Download/Copy and Paste Rules:
Download or copy and paste the .cursor/rules directory into your project. Next, copy and paste the contents of the .cursorrules file in your project root and paste it into your cursor settings rules (Cursor Settings -> Rules -> User Rules). You should also copy and paste the .cursorrules file into your main directory as well.
- Customize Task Workflow (Optional): Once the cursor rules are implemented, restart Cursor and proceed to ask it to create tasks for you based on whatever project idea you may have.
Below is a complete example of an mcp.json file that configures both DevContext and Context7 MCP servers:
{
"mcpServers": {
"devcontext": {
"command": "npx",
"args": ["-y", "devcontext@latest"],
"enabled": true,
"env": {
"TURSO_DATABASE_URL": "libsql://your-project-db-name.turso.io",
"TURSO_AUTH_TOKEN": "your_turso_auth_token_here"
}
},
"context7": {
"command": "npx",
"args": ["-y", "@upstash/context7-mcp@latest"]
}
}
}| Parameter | Description | Default Value |
|---|---|---|
TURSO_DATABASE_URL |
URL of your TursoDB instance | None (Required) |
TURSO_AUTH_TOKEN |
Authentication token for TursoDB | None (Required) |
- System Overview
- Core Components
- MCP Tools
- Data Architecture
- Technical Specifications
- Performance Considerations
- Security
- Troubleshooting
- License
DevContext is a state-of-the-art context management system for software development, implementing the Model Context Protocol (MCP) with advanced context retrieval techniques.
The system operates as a Node.js application with a modular JavaScript codebase, bundled into a single .js file using esbuild. It leverages TursoDB (or similar SQL database) as the persistent store for all context, metadata, and optional logs for a specific project.
Key Differentiators:
- Non-Vector Retrieval: Context retrieval uses sophisticated keyword analysis, relationship graphs, and structured metadata instead of vector embeddings
- Project-Centric Design: Each server instance is dedicated to a single project, simplifying data management
- Minimal Dependencies: Restricted to core essentials - MCP SDK, TursoDB client, and lightweight AST parsing
- Hierarchical Understanding: Context is understood from repository structure down to function/variable levels
- Intelligent Context Prioritization: Multi-factor relevance scoring based on recency, importance, relationships, and developer focus
- Language-aware tokenization with specialized handling for JavaScript/TypeScript, Python, Java, C#, Ruby, and Go
- Keyword extraction with language-specific weighting
- Semantic boundary respecting n-grams
- Language-specific idiom detection
- Code entity indexing and relationship tracking
- Conversation topic segmentation and purpose detection
- Timeline event recording and milestone snapshots
- Focus area prediction based on developer activity
- Code pattern identification and storage
- Automatic pattern learning from examples
- Cross-session pattern promotion
- Design pattern detection
- Query intent prediction
- Multi-factor context prioritization
- Token budget management
- Context integration across topic shifts
DevContext implements the following MCP tools that can be invoked by Cursor IDE:
Initializes a new conversation session with comprehensive project context.
When to use: At the beginning of every conversation, exactly once.
Key parameters:
-
initialQuery: The user's first message or question -
contextDepth: Minimal, standard, or comprehensive context depth -
includeArchitecture: Whether to include architectural context -
focusHint: Optional focus on specific code entity
Returns: Conversation ID and initial context summary
Updates the active context with new messages and code changes.
When to use: After code changes or new messages are exchanged.
Key parameters:
-
conversationId: ID from initialize_conversation_context -
newMessages: New messages exchanged since last update -
codeChanges: Code files created or modified -
preserveContextOnTopicShift: Whether to maintain context during topic changes
Returns: Updated focus and context continuity information
Retrieves context snippets relevant to a specific query.
When to use: When specific project context is needed.
Key parameters:
-
conversationId: ID from initialize_conversation_context -
query: Specific question about the project -
constraints: Optional filters for entity types, file paths, etc. -
weightingStrategy: How to prioritize results
Returns: Relevant context snippets with explanations
Records significant development milestones for future reference.
When to use: After completing important features, fixing critical bugs, or making architectural decisions.
Key parameters:
-
conversationId: ID from initialize_conversation_context -
name: Short, descriptive milestone name -
description: Detailed explanation -
milestoneCategory: Category (feature, bug fix, refactoring, etc.) -
assessImpact: Whether to analyze impact
Returns: Milestone ID and impact assessment
Concludes a conversation, extracting learnings and suggesting next steps.
When to use: At the end of every conversation, exactly once.
Key parameters:
-
conversationId: ID from initialize_conversation_context -
extractLearnings: Whether to identify and extract learnings -
promotePatterns: Whether to promote patterns to global repository -
generateNextSteps: Whether to suggest follow-up actions
Returns: Conversation summary, extracted learnings, and next steps
DevContext uses a SQL database (TursoDB) with the following core tables:
- code_entities: Stores indexed code from files, functions, classes, etc.
- entity_keywords: Maps keywords to code entities for search
- code_relationships: Tracks relationships between code entities
- conversation_history: Stores conversation messages
- conversation_topics: Segments conversations into coherent topics
- timeline_events: Records significant development events
- project_patterns: Stores identified code patterns
- focus_areas: Tracks developer attention and intention
The database schema is automatically created and maintained by the server.
- Node.js: Version 18.0.0 or higher required
- Database: TursoDB (or compatible SQLite)
- Bundling: ESBuild for single-file deployment
- Protocol: Model Context Protocol via @modelcontextprotocol/sdk
- Parsing: Lightweight JavaScript AST parsing (acorn)
- Operating Systems: Cross-platform (Windows, macOS, Linux)
DevContext is optimized for performance with:
- Efficient SQL queries with proper indexing
- In-memory caching for frequently accessed data
- Incremental updates to minimize processing
- Asynchronous operations for non-blocking execution
- Adaptive context retrieval based on token budget
- Scheduled background tasks during idle periods
For large codebases (>100,000 LOC), initial indexing may take several minutes, but subsequent operations remain fast and responsive.
- Isolated Database: Each project uses a dedicated database instance
- Secure Credentials: TursoDB credentials managed via environment variables
- Input Validation: All inputs validated with Zod schemas
- Parameterized Queries: SQL injection protection
- No External APIs: All processing happens locally
Common issues and solutions:
- Connection Errors: Verify TursoDB credentials and database URL
- Slow Initial Startup: Normal for large codebases; subsequent startups are faster
- Missing Context: Check token budget; increase if necessary
- Tool Errors: Ensure proper conversation ID is being passed between tools
- Performance Issues: Consider reducing scope of indexed files or increasing cache size
This project is licensed under the MIT License - see the LICENSE file for details.
DevContext: Continuous Context for Continuous Progress