Glif MCP Server

Official

Generate images, video, and audio with Glif's media-generation agent.

How to install ↓ Compare GitHub source repository ↗

Category: Other
Stars: ★ 213 Very popular
Transport: Streamable HTTP
Runtime: Hosted (no local runtime)
Credentials: API key / credential required
License: MIT
Last commit: 3 days ago
Tools: 2

49FMRS · D

The Glif MCP server is a hosted service offering media generation with support for many clients. It provides intuitive tools like compose_project to simplify multi-step generation, but its main drawbacks are reliance on the hosted platform and credit consumption.

Strongest · Documentation 14/20 Weakest · Reliability 2/20

Reliability

2/20

Security and permissions

9/20

Maintenance

11/20

Documentation

14/20

Setup experience

13/20

Why each score

Reliability 2/20: The repository contains only a server manifest and README; the actual server implementation is hosted on glif.app and not included. There are no tests, no CI workflows, and no way to verify the MCP initialization, tools/list handshake, or whether the declared tools match the live endpoint. The README itself defers to tools/list for the authoritative tool set, creating a consistency gap. All of this is deducted, leaving only a plausible but unverifiable happy path.
Security and permissions 9/20: The manifest and README describe OAuth 2.1 with dynamic client registration and no API keys, which is a strong credential model. The README discloses that generation consumes credits and that the server can run code/search the web. No install examples contain real tokens. However, without the server source code we cannot verify least-privilege scoping, data-flow boundaries, confirmation prompts for paid or side-effecting operations, or the actual OAuth implementation. Those missing verifications cap the score.
Maintenance 11/20: The repository is not archived, has an MIT license, and shows zero open issues. The manifest is versioned at 1.0.1, indicating some maintenance. However, the repository is only registry metadata, so there is little evidence of sustained development, release cadence, or security-response process in the supplied data. No commit history or CI is visible in this static snapshot.
Documentation 14/20: The README is well structured, covering the endpoint, OAuth auth flow, install snippets for Claude, Claude Code, Cursor, VS Code, ChatGPT, Codex, and any client, plus links to pricing, docs, and Discord. It explains the job/polling pattern and the credit model. It is a documentation-only repository, so docs are the primary deliverable and are quite good. Deductions: tool parameters are not documented, limits/troubleshooting are absent, and the README delegates the authoritative tool list to tools/list rather than listing it.
Setup experience 13/20: Because the server is hosted, setup is only a matter of adding the HTTPS endpoint to a client. The README provides one-click/badge installs and copy-paste commands for many clients, plus a generic JSON config and clear OAuth sign-in expectation. That is a clear, few-step path. However, static calibration caps setup at 15 without verifiable execution evidence in the repo (no CI or tests), and the actual connect flow cannot be validated statically, so a modest deduction is applied.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

Users who want to generate media content from MCP clients
Users who prefer OAuth authentication without API keys
Developers who need automated multi-step media generation pipelines

Not for

Users who need a locally-run server and don't want to rely on a hosted service
Users who require fully open-source code
Users who need high-volume generation without spending credits

Required permissions

Access to Glif account (via OAuth)
Generation consumes account credits
Read-only tools are free

Risks and side effects

Generation consumes credits, potentially incurring costs
Dependency on hosted service availability and network connectivity
OAuth login may expose account access

Setup

Before you start

Runtime：Hosted (no local runtime)

Claude (web/desktop): Settings → Connectors → Add custom connector, paste https://glif.app/api/mcp. Claude Code: run claude mcp add --scope user --transport http glif "https://glif.app/api/mcp". Cursor: use the one-click install link or add to .cursor/mcp.json. VS Code: run code --add-mcp '{"name":"glif","type":"http","url":"https://glif.app/api/mcp"}'. ChatGPT: enable developer mode, then Settings → Apps & Connectors → Add new connector, paste URL and pick OAuth. Codex: run codex mcp add glif --url "https://glif.app/api/mcp" and codex mcp login glif. Other clients: use the provided MCP config. Your client opens a browser OAuth sign-in on first connect; approve to link your Glif account.

.cursor/mcp.json

{"mcpServers":{"glif":{"url":"https://glif.app/api/mcp","transport":"http"}}}

Shown for Cursor. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "glif": {
      "url": "https://glif.app/api/mcp",
      "transport": "http"
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add --transport http glif https://glif.app/api/mcp

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After connecting to https://glif.app/api/mcp, the tool list should include compose_project and get_job_status; a browser OAuth sign-in appears on first connect — approve it, then the tools are callable.

Troubleshooting

Check network connectivity and access to https://glif.app/api/mcp
Ensure OAuth login succeeded and account is linked
Refer to https://glif.app/llms.txt for agent docs
Monitor credit balance to avoid unexpected charges

Things to try

Once connected, you can ask your AI assistant things like:

Generate a cyberpunk-style night city image with Glif
Create a 5-second video and check its job status
Transcribe this audio clip to text
Generate three style variations of the same image prompt

Tools 2

compose_project writes

Describe what you want in plain language, including variations, and Glif picks models and chains steps, returning a job_id.

get_job_status read-only

Poll for the media output, which returns resource_link blocks pointing to CDN URLs.

Use cases

Generate images, video, audio

Multi-step media creation workflows

Transcription, HTML rendering, web search, code execution

Supported clients

Claude

Cursor

VS Code

ChatGPT

Codex

Listed from the project's documentation, not tested by this site.

Overview

Glif is a media-generation agent: generate images, video, and audio, transcribe, render HTML, search the web, run code, and chain multi-step media operations — from any MCP client. Endpoint: https://glif.app/api/mcp (Streamable HTTP, JSON-RPC 2.0). Auth: OAuth 2.1 with dynamic client registration — sign in with your Glif account, no API keys. This repo holds the registry metadata for the hosted server (see server.json); the server itself is part of the glif.app platform and is not open source. Note: The old locally-run stdio server (npm @glifxyz/glif-mcp-server) is deprecated — code is on the legacy-local-server branch.

Similar servers

DuckDuckGo Search MCP Server 74 · B

MCP server providing web search and content fetching via DuckDuckGo

★ 1.5k · Tools 2 Compare with this →

Basic Memory 72 · B

Your AI never forgets again.

★ 4.1k · Tools 21 Compare with this →

Telegram MCP Server 70 · B

Telegram MCP server powered by Telethon to let MCP clients read chats, manage groups, and send/modify messages, media, contacts, and settings.

★ 1.8k · Tools 76 Compare with this →

Hostinger API MCP Server 70 · B

Official MCP server for Hostinger API.

★ 160 · Tools 100 Compare with this →

Compare with other MCP servers

Source revision 0c90bf7dbc9f Data synced 2026-10-12 Read the FMRS scoring method