Agent LSP

v0.11.2Developer Toolsstable

MCP server that keeps language server sessions warm and routes multiple languages through one process. Agents get persistent cross-file awareness, speculative execution (simulate edits before writing to disk), and 20 skills that encode correct multi-

agentskillsai-agentsai-toolingclaudeclaude-code
Share:
33
Stars
0
Downloads
0
Weekly
0/5

What is Agent LSP?

Agent LSP is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to mcp server that keeps language server sessions warm and routes multiple languages through one process. agents get persistent cross-file awareness, speculative execution (simulate edits before writing ...

MCP server that keeps language server sessions warm and routes multiple languages through one process. Agents get persistent cross-file awareness, speculative execution (simulate edits before writing to disk), and 20 skills that encode correct multi-

This server falls under the Developer Tools and Coding Agents categories on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • MCP server that keeps language server sessions warm and rout

Use Cases

Maintain persistent language server sessions for agents.
Route multiple languages through single process.
Enable cross-file awareness and speculative code execution.
LicenseMIT License
Languagego
Versionv0.11.2
UpdatedMay 21, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

NPM

npx -y @blackwell-systems/agent-lsp

Manual Installation

npx -y @blackwell-systems/agent-lsp

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Agent LSP

Agent LSP is an MCP server written in Go that keeps Language Server Protocol sessions warm across an entire project and routes 30+ programming languages through a single persistent process. It gives AI coding agents durable cross-file symbol awareness, speculative execution (preview changes in memory before writing to disk), blast-radius analysis, and 65 fine-grained tools that encode correct multi-step refactoring workflows. It is designed specifically for coding agents like Claude Code to dramatically reduce hallucinated edits by grounding every change in real LSP diagnostics.

Prerequisites

  • Go 1.21+ (if installing via go install) or npm/Homebrew for binary installs
  • At least one language server installed for the languages you develop in (e.g. gopls for Go, pyright for Python, typescript-language-server for TypeScript)
  • An MCP client such as Claude Desktop or Claude Code
  • Docker (optional, for the containerized stdio or HTTP modes)
1

Install agent-lsp

Choose the installation method that matches your environment. The install script is the fastest cross-platform option; Homebrew is recommended on macOS.

# macOS via Homebrew
brew install blackwell-systems/tap/agent-lsp

# Any platform via install script
curl -fsSL https://raw.githubusercontent.com/blackwell-systems/agent-lsp/main/install.sh | sh

# npm global install
npm install -g @blackwell-systems/agent-lsp
2

Run interactive setup

Run 'agent-lsp init' to configure which language servers to use and generate the MCP configuration block. Use --non-interactive for CI or scripted environments.

agent-lsp init
3

Configure your MCP client

Add the generated configuration to your MCP client. The args array specifies which language and language server binary to use (e.g. 'go:gopls' for Go).

{
  "mcpServers": {
    "agent-lsp": {
      "command": "npx",
      "args": ["-y", "@blackwell-systems/agent-lsp"]
    }
  }
}
4

Start a language server session

In your AI assistant, instruct it to start an LSP session for your project. The start_lsp tool accepts the root directory of your workspace.

5

Use speculative execution before editing

Before applying code changes, ask the agent to preview the edit with preview_edit or simulate_chain to see diagnostic deltas without touching disk. Only call safe_apply_edit once diagnostics are clean.

Agent LSP Examples

Client configuration

Use this JSON block in claude_desktop_config.json. For Docker stdio mode with Go/gopls, see the alternate form below.

{
  "mcpServers": {
    "agent-lsp": {
      "command": "npx",
      "args": ["-y", "@blackwell-systems/agent-lsp"]
    }
  }
}

Prompts to try

These prompts exercise the LSP-grounded tools for cross-file intelligence, safe refactoring, and code quality audits.

- "Start an LSP session for /my/project and show me all callers of the processOrder function"
- "What is the blast radius if I rename the UserService interface? Show me every file that imports it"
- "Preview renaming the variable 'ctx' to 'requestCtx' in handler.go and report any diagnostics"
- "Run a 12-check code quality audit on the auth package and list the findings"
- "Safely delete the deprecated fetchLegacyData function and verify no errors are introduced"

Troubleshooting Agent LSP

Language server binary not found when starting LSP session

Install the appropriate language server for your language (e.g. 'go install golang.org/x/tools/gopls@latest' for Go, 'npm install -g typescript-language-server typescript' for TypeScript) and ensure it is on your PATH.

AGENT_LSP_OUTPUT_FORMAT causes unreadable output in the MCP client

Set AGENT_LSP_OUTPUT_FORMAT=json in the env block of your MCP config to switch from Graph Compact Format to standard JSON, which most clients handle more reliably.

HTTP mode returns 401 Unauthorized

Set the AGENT_LSP_TOKEN environment variable to a secret string and pass the same value as a Bearer token in your client's Authorization header.

Frequently Asked Questions about Agent LSP

What is Agent LSP?

Agent LSP is a Model Context Protocol (MCP) server that mcp server that keeps language server sessions warm and routes multiple languages through one process. agents get persistent cross-file awareness, speculative execution (simulate edits before writing to disk), and 20 skills that encode correct multi- It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Agent LSP?

Install via npm with the command: npx -y @blackwell-systems/agent-lsp. Then add the server configuration to your AI client's JSON config file (e.g., claude_desktop_config.json or .cursor/mcp.json).

Which AI clients work with Agent LSP?

Agent LSP works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Agent LSP free to use?

Yes, Agent LSP is open source and available under the MIT License license. You can use it freely in both personal and commercial projects.

Browse More Developer Tools MCP Servers

Explore all developer tools servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "agent-lsp": { "command": "npx", "args": ["-y", "@blackwell-systems/agent-lsp"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Agent LSP?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides