This repository documents the Model Context Protocol (MCP) servers configured for Kilo IDE to enhance AI-assisted development.
Kilo IDE is configured with 8 MCP servers that provide various tools for code management, browser debugging, cloud services, and memory.
- Type: Remote
- URL: https://mcp.grep.app
- Purpose: Search code across GitHub repositories
- Usage: Find code snippets, functions, and implementations
- Type: Local
- Command:
npx -y chrome-devtools-mcp@latest --executablePath "C:\Program Files (x86)\Microsoft\Edge\Application\msedge.exe" - Purpose: Browser automation and debugging
- Features:
- Navigate pages
- Take screenshots
- Execute JavaScript
- Monitor network requests
- Fill forms and interact with elements
- Type: Local
- Command:
npx -y @modelcontextprotocol/server-github - Environment:
GITHUB_PERSONAL_ACCESS_TOKEN - Purpose: Full GitHub repository management
- Features:
- Create/update files and repositories
- Manage issues and pull requests
- Search code and issues
- Fork repositories
- Create branches and commits
- Type: Local
- Command:
npx -y @cloudflare/mcp-server-cloudflare run - Purpose: Manage Cloudflare resources
- Features:
- Deploy Workers
- Manage KV namespaces
- Configure D1 databases
- Set up R2 storage
- Manage environment variables
- Type: Local
- Command:
npx -y @modelcontextprotocol/server-filesystem "C:\Users\Administrator\Documents\Codex" "C:\tmp" - Purpose: Safe read/write access to project files
- Features:
- Read files and directories
- Write and update files
- Create/delete files
- Restricted to configured directories
- Type: Local
- Command:
npx -y html-extractor-mcp - Purpose: Convert web pages to plain text
- Features:
- Fetch URLs
- Extract text content
- Clean HTML tags
- Preserve structure
- Type: Local
- Command:
npx -y @modelcontextprotocol/server-sequential-thinking - Purpose: Step-by-step complex reasoning
- Features:
- Break down complex problems
- Revise and refine thoughts
- Branch into alternative reasoning paths
- Generate and verify hypotheses
- Type: Remote
- URL: https://mcp.mem0.ai/mcp/
- Headers:
Authorization: Token m0-... - Purpose: Persistent vector memory for AI assistants
- Features:
- Semantic search across memories
- Multi-tenant support
- Entity relationship tracking
- Cross-device synchronization
- REST API and MCP protocol
Save text or conversation history for a user/agent.
Semantic search across memories with filters.
List or retrieve specific memories.
Update or delete individual memories.
List users, agents, apps, or runs stored in Mem0.
The MCP servers are configured in:
~/.config/kilo/kilo.jsonc
The GitHub MCP server provides 26+ tools:
- Repository Management: Create/update repos, manage settings
- File Operations: Create/update/delete files in repositories
- Branch Management: Create/switch/delete branches
- Issues: Create, list, update, and manage issues
- Pull Requests: Create, manage, review, and merge PRs
- Code Search: Search code across repositories
- Commits: List and inspect commits
- Collaboration: Manage comments, reviews, and reactions
The Chrome DevTools MCP server provides 27+ tools:
navigate- Navigate to URLback/forward/reload- Browser history
click- Click elements by selector or coordinatestype- Type text into inputsfill_form- Fill multiple form fieldsscroll- Scroll page or elementsdrag_drop- Drag and drop elements
get_text- Extract visible textget_html- Get element HTMLevaluate- Execute JavaScriptquery_selector- Find elements by CSS selector
screenshot- Capture screenshotsget_page_info- Get page metadata
network_enable- Start capturing network requestsnetwork_get_log- Get captured requestsget_response_body- Get response content
performance_trace- Record performance traceslighthouse_audit- Run Lighthouse audits
device_emulation- Simulate devicesviewport- Set viewport sizegeolocation- Set geolocation
- Research: Use
gh_grepto find code examples - Context: Use
fetchto read documentation - Implementation: Use
filesto read/write code - Testing: Use
chrome-devtoolsto test in browser - Deployment: Use
cloudflareto deploy Workers - Memory: Use
mem0to remember project context
- AI searches for similar patterns with
gh_grep - AI reads current code with
files - AI uses
sequential-thinkingto analyze - AI suggests improvements
- AI applies changes with
files - AI tests with
chrome-devtools
- Conversations are saved to
mem0cloud - Memories are accessible from any device
- Semantic search finds relevant context
- Entity relationships track project evolution
The files MCP server is restricted to:
C:\Users\Administrator\Documents\CodexC:\tmp
This prevents access to sensitive system files.
The Cloudflare MCP uses OAuth tokens with limited permissions:
- Workers: read/write
- KV: read/write
- D1: read/write
- R2: read/write
- Analytics: read-only
GitHub operations use personal access tokens with repository-specific permissions.
- Check Kilo configuration file syntax
- Verify all commands are installed globally
- Check environment variables for tokens
- Restart Kilo IDE
- Verify
MEMORY_FILE_PATHis writable - Check Mem0 API key is valid
- Test Mem0 connection with
mem0CLI
- Run
wrangler whoamito verify login - Check token permissions
- Verify account has required resources
- Install the MCP server package
- Add configuration to
kilo.jsonc - Specify type (local/remote)
- Define command and arguments
- Set environment variables if needed
- Restart Kilo
- Least Privilege: Restrict filesystem and cloud access
- Token Management: Use environment variables for secrets
- Memory Hygiene: Regular review of stored memories
- Code Reviews: Always review AI-generated code
- Testing: Test AI modifications before committing
- Model Context Protocol
- MCP Servers Registry
- Awesome MCP Servers
- Mem0 Platform
- Cloudflare Workers
- Chrome DevTools Protocol
This configuration is provided as-is for Kilo IDE users. Individual MCP servers have their own licenses.