About previewglow
previewglow is a local bridge to Shareglow's hosted MCP server for inspecting and managing link previews and Open Graph images. It serves the upstream tools locally and forwards calls and results without retaining preview data.
Key features
- 01Scan pages for preview metadata and issues
- 02Render Open Graph images from templates
- 03Create templates from website branding
- 04Create short links with preview variants
- 05View click analytics by several dimensions
- 06Refresh previews on social platforms
Use cases
- 01Check how a public page appears on social platforms
- 02Generate an Open Graph image using a workspace template
- 03Create branded templates from a website
- 04Review short-link clicks or refresh changed previews
MCP support
Exposes Shareglow's tools through a local MCP server and forwards tool calls to its hosted MCP endpoint. Upstream results are returned unchanged.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
7Saved tool list · updated 1w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| url | string | Yes | The public web page URL to scan for link-preview (Open Graph / Twitter Card) metadata. |
Details
From the project READMEpreviewglow
previewglow is a small local bridge to the hosted Shareglow MCP server, which looks after link previews and Open Graph images. Point it at a public page and it reports the page's preview metadata, the per-platform previews, and any problems it finds, with suggestions for fixing them. It can render a 1200x630 OG image from one of the workspace's templates, design a fresh template from a site's own brand, mint short links that carry their own previews, report how those links are being clicked, and force the social platforms to re-fetch a page whose tags have just changed.
The bridge keeps no preview data of its own. It serves the upstream server's tool list locally, exactly as captured from the live endpoint, and forwards each tool call upstream, returning the result untouched. Everything here acts on the caller's own Shareglow workspace or on public pages the caller names; there is no sending, payment, or messaging tool in the surface.
Tools
scan_url— scan a public page and report its link-preview (Open Graph / Twitter Card) metadata, previews per platform, and issues with fix suggestions. Required:url.generate_og_image— render a 1200x630 Open Graph image from a template. Required:templateId. Optional:params(text values such as title and subtitle) andexp(Unix seconds; makes the signed URL expire, paid plans). Returns the permanent image URL and whether it was already cached.create_template_from_url— use AI to design a new OG image template from a website's brand (logo, colors, fonts). Required:url. Returns the chosen layoutId, the variables to pass, and the extracted brand kit.list_templates— list the OG image templates the workspace can render (built-in and custom) with id, name, version and dimensions. Takes no arguments; call it beforegenerate_og_imageto pick a templateId.create_short_link— create a short link with its own link preview. Required:destination. Optional:slugand up to three A/Bvariants. Returns id, shortUrl and variants.get_link_stats— get click analytics for a short link: totals plus breakdowns by day, country, device, browser and referrer. Required:linkId. Optional:from/todates (YYYY-MM-DD).refresh_preview— force social platforms to re-fetch a page's link preview after its OG tags or image changed. Required:url.
Configuration
SHAREGLOW_MCP_URL(optional) — override the upstream endpoint. Defaults tohttps://mcp.getshareglow.com/mcp. (The override hook is this rebuild's own labelled reconstruction.)SHAREGLOW_API_KEY(optional) — when set, every upstream request carries anAuthorization: Bearer <key>header, for the caller's own Shareglow key (ansk_live_...key created in the caller's Shareglow dashboard). When it is not set, no key header is sent. (The header name and the Bearer template are pinned by the original's registry record; forwarding the key from this env var is a labelled reconstruction — the original surface was captured keyless.)
Forwarding mechanics
On the first tool call the client opens a streamable
HTTP session with the upstream: a JSON-RPC
initialize request (protocolVersion 2025-03-26,
clientInfo name previewglow), then a
notifications/initialized notification, then the
tools/call request carrying the caller's arguments
unchanged. When the upstream answers initialize
with an Mcp-Session-Id header, that value rides on
every later request of the session, which is reused
for the life of the process. Replies may arrive as
plain JSON or as server-sent events; both are read.
Successful upstream results pass through verbatim,
including any isError flag the upstream sets.
Failures — a JSON-RPC error, a non-2xx status, an
unreadable response, or a request that cannot be
completed — come back as an isError result whose
text starts with previewglow error: (this rendering
is this rebuild's own labelled reconstruction).
CLI
Each tool is also a subcommand taking one JSON argument string:
npx @reneturcios/previewglow scan_url '{"url":"https://example.com/landing"}'
npx @reneturcios/previewglow get_link_stats '{"linkId":"lnk_abc-123"}'
npx @reneturcios/previewglow --version
Exit codes: 0 on success, 1 on an isError
result, 2 on a usage error.
MCP server
npx -y @reneturcios/previewglow mcp
MCP client config:
{
"mcpServers": {
"previewglow": {
"command": "npx",
"args": ["-y", "@reneturcios/previewglow", "mcp"]
}
}
}
Docker
docker run -i ghcr.io/cloudcorprecords/previewglow
docker run --rm ghcr.io/cloudcorprecords/previewglow --version
The container's default command is mcp, so the first
form is a stdio MCP server; any other arguments reach
the CLI normally.
Development
npm install
npm run build
npm test
npm run dev
The test suite runs entirely against a stubbed JSON-RPC upstream — no live network calls.
License
MIT — see LICENSE.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.