lastseen-mcp
Checks whether software dependencies (npm packages, GitHub Actions, MCP servers, Docker images) are alive, dormant, abandoned, archived, or deleted using dated observations from lastseen.dev.
README
lastseen-mcp
<!-- MCP registry ownership --> mcp-name: dev.lastseen/mortality
Is this dependency dead? Dated observations of whether an npm package, GitHub Action, MCP server, or Docker image is alive, dormant, abandoned, archived, or deleted — published by lastseen.dev. Free, no API key, zero telemetry.
There are two ways in, and they answer the same question from the same dated data:
- Call the free HTTP API directly — the durable, machine-native path (below). An agent that keeps a dependency list can re-check it on every build.
- Run it as an MCP server — a thin stdio wrapper over that same API, for MCP clients.
1. The free API (call it directly)
No key. No signup. 60 requests/minute, anonymous GET. Every response is a dated
observation with its source — not a score, not advice. Base URL https://lastseen.dev.
# one component (owner/repo)
curl https://lastseen.dev/api/v1/entity/actions/checkout
# everything held under a GitHub owner/org
curl https://lastseen.dev/api/v1/org/actions
# a whole manifest at once (names-only body)
curl -X POST https://lastseen.dev/api/v1/manifest \
-H 'content-type: application/json' \
-d '{"channel":"api","components":[{"name":"actions/checkout","kind":"github-action"}]}'
# self-describing API docs (JSON)
curl https://lastseen.dev/api/v1/docs
A single-component response carries the latest_state, the dated series behind it, the
source_of_record, and interval-censoring notes — enough to re-check any verdict against GitHub
yourself. This is the recurring path: wire GET /api/v1/entity/{owner}/{repo} into CI and a
dead dependency shows up on the build that introduces it, not months later.
Contract
- Lookups, not advice. Every answer is a dated observation with its source. Never "use X instead", never a composite score, never a ranking by preference, never urgency language.
- Fails closed. An unknown or fabricated subject returns a clean
not_observed— never a fabricated result. A network/store error degrades (try later); it is never read as "dead". - Absence ≠ gone. No record means "not checked", never "deleted". Only an authoritative HTTP 404 is a deletion.
States
alive (≤180d since last commit) · dormant (180–365d) · abandoned (>365d) ·
archived (repo archived) · deleted (authoritative HTTP 404) · eol (declared end-of-life) ·
unknown-stale (the observation the verdict rests on is older than the 180-day threshold and was
not re-verified — no state asserted; not a death) · not_observed (403 / 429 / timeout / not
held — not dead).
2. Run it as an MCP server
A pure-standard-library stdio wrapper over the API above — no mcp package, no requests,
no FastAPI/Starlette, nothing that can conflict with a host app. It speaks JSON-RPC 2.0 over
stdio and calls the free API over urllib.
Install
pipx install lastseen-mcp
# or run without installing:
uvx lastseen-mcp
Requires Python ≥ 3.10. No dependencies.
Add to an MCP client
{
"mcpServers": {
"lastseen": { "command": "lastseen-mcp" }
}
}
(If you use uvx, set "command": "uvx", "args": ["lastseen-mcp"].)
Tools
| Tool | What it does |
|---|---|
check_entity(slug, live?) |
Dated survival series for one owner/repo (e.g. actions/checkout). live=true attaches a live GitHub read (absence ≠ gone). |
check_org(name) |
Dated states for every held component under a GitHub owner/org. |
check_manifest(paste) |
Paste a GitHub Actions workflow (uses: refs) or owner/repo lines; get the dated state per component. Unheld components are not_observed. |
survival_profile(category) |
Category ranking — not available over the free API; returns an honest not-supported result pointing to the per-entity/org/manifest lookups and the CC-BY dumps. |
check_manifest looks up at most 30 parsed components per call (free-tier rate bound); any
remainder is reported as n_truncated.
Privacy
Zero telemetry. The client and the API store nothing and transmit no identifier about you, your org, or the subjects you look up — anonymous GETs only.
Licensing
- Code: MIT (see
LICENSE). - Data: the mortality/survival observations are published by lastseen.dev under CC-BY-4.0 (attribution required). The two licenses are independent.
Links
- Service & data: https://lastseen.dev
- API docs (self-describing JSON): https://lastseen.dev/api/v1/docs
- MCP registry entry:
dev.lastseen/mortality
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.