Hlido

agent-lsp

Coding · tested 2026-08-27 · by the Hlido desk, not the vendor

In short: A stateful MCP bridge to real language servers — 66 tools across 30 languages whose support is verified by running the actual language server in CI, not listed in a config file.

4 PASS · 1 FAIL of 5 public-surface claims

Visit agent-lsp → scorecard.json

Quick answer

agent-lsp scores 79/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-08-27). STEADY (79) because agent-lsp makes a falsifiable verification claim and describes the mechanism behind it — 30 real language servers run against real fixtures in CI on every push — which is a materially stronger evident Pricing: Usage-based.

Hlido Editor · 2026-08-27

agent-lsp identifies the failure mode precisely: coding agents make wrong changes because they cannot see who calls a function, what a rename breaks, or whether the build still passes — and language servers already know all of that. Existing MCP-to-LSP bridges either cold-start on every request or expose raw LSP primitives that agents then use incorrectly. agent-lsp's answer is a stateful runtime that indexes the workspace once and keeps the index warm across files, packages and repositories, plus a skill layer that encodes correct multi-step operations so they actually complete rather than stalling halfway. One process routes .go to gopls, .ts to typescript-language-server and .py to pyright with no reconfiguration when you switch projects, and it ships as a single Go binary. The claim we rate highest is the verification one, because it is the rare kind that is falsifiable: 'Every other MCP-LSP implementation lists supported languages in a config file. None of them run the actual language server in CI to verify it works.' agent-lsp states its CI runs 30 real language servers against real fixture codebases on every push, and names them — Go, Python, TypeScript, Rust, Java, C, C++, C#, Ruby, PHP, Kotlin, Swift, Scala, Zig, Lua, Elixir, Gleam, Clojure, Dart, Terraform, Nix, Prisma, SQL, MongoDB and more. 'When we say works with gopls, that's a verified, automated claim, not a hope' is exactly the posture we want vendors to take, and it is the main reason this scores where it does. The counterweight is scale: 116 stars and v0.18.0, from a single organisation, for a component that would sit in the critical path of every code edit an agent makes. The comparative claim about other implementations is the vendor's own characterisation of competitors, and the token-savings and speculative-execution claims are documented but not measured here.

Why STEADY

STEADY (79) because agent-lsp makes a falsifiable verification claim and describes the mechanism behind it — 30 real language servers run against real fixtures in CI on every push — which is a materially stronger evidentiary posture than the config-file support lists it criticises, and the stateful warm-index design addresses a real, specific agent failure. Held below the top of the band because adoption is small (116 stars, v0.18.0), the CI claim is stated on the vendor's own documentation rather than independently confirmed here, and its characterisation of competing implementations is vendor-authored.

Public-surface checklist

What we saw

1 screenshot captured by the Hlido engine during the reviewed run (run-7ccf6419a44ea8e6-www-agent-lsp-com). Our own captures — not vendor marketing material.

agent-lsp — run screenshot 1 (home.png)
home.png

What it does well

What it fails at

Best for

  • Coding agents that need real call-graph, rename-impact and build-validity answers rather than inference from text
  • Polyglot monorepos where one warm index across many languages beats per-project reconfiguration
  • Teams that specifically value a vendor verifying its support matrix in CI rather than declaring it

Not recommended for

  • Teams that require a mature, widely adopted dependency in the critical path
  • Languages outside the 30 CI-verified set, where the verification argument does not apply
  • Buyers who need a support agreement or licensing commitment before adopting

Pricing & access

Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-08-27.

Compared to

Agent relevance

CLI MCP Behavioral-testable

Agentic-Commerce Readiness 64/100 · INTEGRABLE

Independent readiness for agent delegation & transaction. How it’s scored · check live

Install the single Go binary and connect it as an MCP server to Claude Code or any MCP host, pointing it at a code root. The agent then calls the 66 language-intelligence tools and the 24 encoded agent workflows; the skill layer sequences multi-step operations (rename, find-references, refactor) so they complete correctly rather than exposing raw LSP primitives the agent has to orchestrate itself.

Agent-friendly score: 9/10

Evidence

scorecard.json · transparency passport · registry · methodology

More: compare agents · best of · developer tools · incident registry

Verdict by Hlido Editor, our automated editorial system · Method: public-surface-tier-1+editorial-narrative-v2 · Methodology version 2026.05 ·

How this page was produced. The scores, claim verdicts and evidence come from automated hands-on testing of the product’s public surface. The written analysis is drafted by an AI system, and pages publish without a person reviewing each one. Hlido publishes this record and answers for it — tell us if anything here is wrong and we will correct it.

Embed this trust badge

Hlido trust score

Live, always-current independent score — free to embed on your site or README. No vendor pays for placement.

Markdown

[![Hlido trust score](https://hlido.eu/badge/blackwell-systems-agent-lsp.svg)](https://hlido.eu/check/?agent=blackwell-systems-agent-lsp)

HTML

<a href="https://hlido.eu/check/?agent=blackwell-systems-agent-lsp"><img src="https://hlido.eu/badge/blackwell-systems-agent-lsp.svg" alt="Hlido trust score"></a>