MegaLens

ai.megalensPublished by megalens.aiOfficial publisherv1.0.0Updated Nov 2, 2026

Code review by AI models from different companies, usually two. Shows where they agree and disagree.

VerifiedStreamable HTTPWeb executableDeveloper ToolsAI & ML

Overview

AI-generated overview

MegaLens runs AI code review through a panel of models from different companies and reports where they agree and disagree.

What you can do
  • Submit code for review by a panel of up to four AI models from different companies, which review it separately.
  • Get back a consolidated result showing where the models agreed and where they disagreed, naming the models that took part.
  • Compare the panel's findings against your own assessment when you send one with the request.
  • Set up and check the connection with helper commands that detect installed tools, write configs, validate the token and show config status.
When to use it
Useful when you want a second opinion on code from several independent models inside the AI coding tool you already use, and want disagreements between them surfaced rather than a single verdict. The first review is free, so it is easy to try before paying.
Requirements
Remote MCP endpoint at megalens.ai, authenticated with an Authorization bearer token obtained from the MegaLens settings page. A MegaLens account is required. The optional setup wizard needs Node.js 18 or later; manual configuration needs nothing. Optionally an OpenRouter key header for bring-your-own-key use.
Before you install
The Authorization bearer token is a secret and must be kept out of shared configs. Reviews are sent to MegaLens and to third-party model providers. Paid use costs money: pay-as-you-go credits are prepaid, or you pay OpenRouter directly when supplying your own key, which needs credit on it. The server reports only and does not change your code.

Installation

In SourceWeft

  1. Open MegaLens in the dashboard and add it to a workspace.
  2. Connect the required credentials. They are stored privately and never shown on public pages.
  3. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/ai.megalens%2Fmcp/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config. The client completes OAuth sign-in on first connect.

{
  "mcpServers": {
    "megalens": {
      "type": "http",
      "url": "https://megalens.ai/api/mcp"
    }
  }
}

README

megalens-mcp

Code review from a panel of AI models, inside the AI tool you already use. MegaLens selects up to four AI models from different companies to review your code, and reports where they agree and where they disagree. On Free, MegaLens selects two reviewing models.

Works with Claude Code, Codex CLI, Cursor, Gemini CLI and Lovable. GitHub Copilot is coming soon.

Your first review is free, with no card.

Quick Start

bash
npx megalens-mcp setup

The setup wizard:

  1. Asks for your MegaLens token (get one here)
  2. Detects installed tools automatically
  3. Writes the correct config for each tool

The wizard does not ask for an OpenRouter key. To use your own key, add one header by hand: see Adding your own OpenRouter key.

Supported Tools

ToolStatusConfig FileAuto-detected
Claude CodeSupported~/.claude.jsonYes
Codex CLISupported~/.codex/config.tomlYes
CursorSupported~/.cursor/mcp.jsonYes
Gemini CLISupported~/.gemini/settings.jsonYes
LovableSupportedNone (a form in Lovable)No, see below
GitHub Copilot (VS Code)Coming soon, not supported yet.vscode/mcp.jsonOnly if that file already exists in the folder you run the wizard from

If no tool is detected, the wizard lets you pick one and creates the config file. The wizard can also write configs for Windsurf and Grok Build. Those are not supported or tested.

Lovable needs no config file and no wizard. Paste the server URL and your token into Lovable's MCP connector form. Steps: megalens.ai/integrations/lovable

Manual Setup

The tools use two config languages and three different top-level keys, which is why the wizard exists.

Keep the ?ide= on each URL. It tells MegaLens which tool the review came from. It does not change which models review your code: the panel is the same whichever tool you call it from.

Claude Code

Run this once. The --scope user flag makes MegaLens available in every project:

bash
claude mcp add --transport http --scope user megalens "https://megalens.ai/api/mcp?ide=claude-code" \  --header "Authorization: Bearer ml_tok_your_token_here"

Or in .mcp.json (project) / ~/.claude.json (user):

json
{  "mcpServers": {    "megalens": {      "type": "http",      "url": "https://megalens.ai/api/mcp?ide=claude-code",      "headers": {        "Authorization": "Bearer ml_tok_your_token_here"      }    }  }}

Cursor

~/.cursor/mcp.json or .cursor/mcp.json. No type field. Cursor does not use one:

json
{  "mcpServers": {    "megalens": {      "url": "https://megalens.ai/api/mcp?ide=cursor",      "headers": {        "Authorization": "Bearer ml_tok_your_token_here"      }    }  }}

Gemini CLI

~/.gemini/settings.json. Gemini CLI uses httpUrl, not url:

json
{  "mcpServers": {    "megalens": {      "httpUrl": "https://megalens.ai/api/mcp?ide=gemini-cli",      "headers": {        "Authorization": "Bearer ml_tok_your_token_here"      }    }  }}

GitHub Copilot in VS Code (coming soon, not supported yet)

GitHub Copilot is coming soon. If you want to try it early: .vscode/mcp.json. The top-level key is servers, not mcpServers. A config written for Claude Code or Cursor will not be read here:

json
{  "servers": {    "megalens": {      "type": "http",      "url": "https://megalens.ai/api/mcp?ide=copilot",      "headers": {        "Authorization": "Bearer ml_tok_your_token_here"      }    }  }}

Codex

~/.codex/config.toml or .codex/config.toml. TOML, not JSON:

toml
[mcp_servers.megalens]url = "https://megalens.ai/api/mcp?ide=codex"http_headers = { "Authorization" = "Bearer ml_tok_your_token_here" }

Adding your own OpenRouter key (free / BYOK)

Add one more header alongside Authorization, in whichever shape your tool uses above. The key needs credit on it. You pay OpenRouter directly; MegaLens is not part of that bill.

"x-megalens-openrouter-key": "sk-or-v1-your-openrouter-key"

In Codex TOML:

toml
http_headers = { "Authorization" = "Bearer ml_tok_your_token_here", "x-megalens-openrouter-key" = "sk-or-v1-your-openrouter-key" }

With Claude Code's command, add --header "x-megalens-openrouter-key: sk-or-v1-your-openrouter-key".

If you skip the key, use pay-as-you-go credits instead: $9 per 1M blended tokens, prepaid from $20.

Commands

CommandDescription
megalens-mcp setupInteractive setup: detect tools, write config
megalens-mcp validateTest your token connectivity
megalens-mcp configShow current config status

How It Works

MegaLens selects up to four AI models from different companies to review your code. They review it separately. A final check reviews their findings, and you get back where they agreed and where they disagreed. Each result names the models that actually took part.

If you send your own assessment with the request, the answer also says what the panel found that yours did not.

The panel is the same whichever tool you call it from. MegaLens reports; it does not change your code. Your own tool stays the one that decides what to do next.

Our private recipe. What stays private: how we choose the models and what each one is asked to do, the numbers that tune each check, and the prompts. That recipe came from months of measured runs and changes as models change. Our results, rules and mistakes are public: megalens.ai/research/ai-code-review-false-positives

Requirements

  • Node.js 18+ (only for the setup wizard; manual setup needs nothing)
  • A MegaLens account: megalens.ai. Your first review is free, with no card

Links

Source: README.md at commit 8bdf563

Tools

0
This server asks you to sign in, so SourceWeft could not read its tools without an account and has not indexed them. What a connected account sees can depend on that account and its permissions.

Version history

1
  1. v1.0.0LatestSep 23, 2026

Score

85% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts