Welcome to the LSP Client ecosystem — a professional-grade toolkit and Agent-native ecosystem built around the Language Server Protocol (LSP).

We provide a complete stack to transform raw language intelligence into high-fidelity context for humans and AI Agents alike.

The industrial-strength foundation. A modern, async-first library for fine-grained LSP control.

• Mixin Architecture: Zero-overhead system for automatic capability negotiation and method safety.
• Async-First: Built with anyio for high-concurrency tool development.
• Production Ready: Full type safety, explicit environment control, and robust error handling.

The isolation layer. Optimized, pre-built Docker images for instant, secure server deployment.

• Zero-Config Isolation: Protect your host environment and ensure reproducible analysis.
• Auto-Updated: Weekly builds to keep you at the bleeding edge of language support.
• Lean & Fast: Multi-stage builds for minimal footprint and rapid startup.

The intelligence layer. Designed for the next generation of AI tools and coding assistants.

• LSAP (Language Server Agent Protocol): A semantic abstraction that transforms granular LSP operations into agent-friendly Markdown snapshots.
• Model Context Protocol (MCP): Standardized integration to expose LSP and LSAP capabilities directly to LLM-powered agents.
• Progressive Disclosure: Engineered to provide high-signal context with minimal noise for autonomous reasoning.

Our ecosystem supports a growing list of language servers, available locally or via our container registry:

Integrate language intelligence into your project using the Python SDK:

import anyio
from lsp_client.clients.pyright import PyrightClient, PyrightContainerServer
from lsp_client.common import Position

async def main():
    # Use Pyright in a container for instant isolation and zero configuration
    async with PyrightClient(server=PyrightContainerServer()) as client:
        definitions = await client.request_definition_locations(
            file_path="app.py",
            position=Position(line=10, character=5)
        )
        for loc in definitions:
            print(f"Found definition: {loc.uri} at {loc.range}")

anyio.run(main)

We are building a robust, open ecosystem for the future of software engineering.

• Contribute a Server: Add a new ContainerFile to our registry.
• Extend the Protocol: Help us evolve LSAP and the Python SDK.
• Build Agents: Use our tools to create the next generation of AI coding assistants.

Check out our Contribution Guidelines to get started.

Built with ❤️ by the LSP Client team.