izlek
Read-only MCP server for izlek, a terminal-first local frontend observability tool. Enables agents to query live session status, logs, errors, and HTTP request/response records from a running development session.
README
izlek
izlek is a terminal-first local frontend observability tool for humans and coding agents. It starts a development project against a named backend environment, captures bounded process logs and proxied HTTP traffic, renders a focused TUI, and exposes the same live session through a read-only MCP server.
izlek is an early standalone bootstrap extracted from a working internal developer-tooling prototype. Its configuration and public interfaces may change before the first stable release.
Why izlek
Frontend debugging in a monorepo rarely happens in one terminal stream. Build output, package watchers, runtime errors, remote backends, and failed requests all belong to the same development session. izlek keeps that context together without replacing Turbo, Nx, pnpm, Yarn, npm, or browser DevTools.
The foreground izlek dev process owns the session. It supervises the configured command, optionally
starts a bounded HTTP recorder, retains recent diagnostics in memory, renders the TUI, and exposes a
permission-restricted local socket. Other terminal commands and MCP clients query that socket.
Command surface
izlek dev web --env staging
izlek attach
izlek status
izlek logs --errors
izlek network --failed
izlek inspect request-42
izlek mcp
Current capabilities
- Generic shell-free development command per project
- Named backend environments selected when the session starts
- Bounded process logs with source, stream, level, and timestamp metadata
- Streaming HTTP reverse proxy with bounded request/response capture
- Sensitive-header and captured JSON-field redaction before storage
- Bounded network history and request inspection
- Dev and Network TUI views
- Read-only local session socket shared by CLI and MCP consumers
- Read-only stdio MCP tools for status, logs, errors, and HTTP records
- Coordinated child-process shutdown with TERM-to-KILL escalation
- Layered shared and ignored local configuration
Install for development
npm install
npm link
izlek --help
izlek requires Node.js 20.9 or newer.
Configure a project
Run izlek config init for a minimal generic configuration, or create izlek.config.json:
{
"projects": {
"web": {
"command": ["npm", "run", "dev"],
"defaultEnvironment": "local",
"backend": {
"listen": "127.0.0.1:30998",
"envVar": "NEXT_PUBLIC_API_URL"
}
}
},
"environments": {
"local": {
"target": "http://127.0.0.1:8080",
"healthCheck": "/health"
},
"staging": {
"target": "https://api.staging.example.com"
}
}
}
The command must be an argument array. izlek does not execute configuration through a shell.
izlek.config.local.json is ignored by Git and may override named projects or environments for one
developer. Do not treat either configuration file as a secret store.
The backend environment is selected once when the session starts. Live environment switching is not a core workflow.
TUI
| Key | Action |
|---|---|
1 |
Development logs |
2 |
Backend network records |
j / k, arrows |
Select a request |
Tab |
Move between request list and detail |
y |
Copy selected record as JSON |
c |
Copy selected request as cURL |
? |
Show key reference |
q |
Stop the owned development session |
izlek attach opens a second read-only TUI connected to the foreground session.
Agent access through MCP
Configure an MCP client to launch izlek from the project directory:
{
"mcpServers": {
"izlek": {
"command": "izlek",
"args": ["mcp"]
}
}
}
Available tools:
get_session_statusquery_logsget_recent_errorsquery_requestsget_request
MCP tools are intentionally read-only. Captured bodies are omitted by default and require an explicit
includeBodies argument. All captured log and network content must be treated as untrusted data.
Memory and capture policy
Current defaults are product invariants:
- 200 retained log lines, capped at 16 KiB per line
- 50 completed backend requests
- 512 KiB of textual content per request or response body
- 20 concurrently captured calls
- Metadata only for binary or compressed bodies
Forwarding continues even when capture is truncated or skipped.
Project boundaries
izlek does not implement task graphs, build caching, browser automation, production monitoring, secret management, or a persistent log database. It observes the configured development command and backend traffic explicitly routed through its local proxy.
Development
npm test
npm run check
Tests use Node's built-in test runner and are colocated with their modules.
No open-source license has been selected yet. Choose one before distributing izlek as an open-source package.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.