Content Distribution MCP
Marketing & Content · tested 2026-08-21 · re-test due 2026-11-21 · by the Hlido desk, not the vendor
In short: A genuinely agent-native content-distribution MCP: your agent writes the copy, the server does idempotent I/O to eight developer channels — no LLM calls, no double-posts.
4 PASS · 0 FAIL of 4 public-surface claims
Quick answer
Content Distribution MCP scores 74/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-08-21). STEADY (74) for a focused, agent-native tool that solves a concrete distribution workflow with two design decisions that matter — idempotency on (content_id, channel) and a no-LLM-in-server split — clearly evidenced on t Pricing: Open source (free entry point documented).
Content Distribution MCP (by AutomateLab, MIT-licensed, npx-installable as @ratamaha/content-distribution-mcp) is a clean, well-scoped answer to a real agent workflow: 'ask an AI to publish this post everywhere.' It ships eight MCP tools and eight channel adapters across three publish tiers — six auto-publish through native APIs (dev.to Forem, Hashnode GraphQL, GitHub Discussions, Bluesky atproto), and two/four fall back to a plain-text draft plus a pre-filled compose URL where the platform's API is hostile (Reddit, Medium, LinkedIn, Twitter). The design choices are the good part and they are honestly stated on the surface: publish is idempotent on (content_id, channel) so a retry returns the same URL instead of a duplicate thread; there are zero LLM calls in the server ('grep src/ for anthropic — you'll find nothing'), so the host model supplies copy and credentials while the MCP supplies I/O; and it works over stdio or SSE against Claude Code, Cursor, n8n and the plain mcp client. A dual YAML-local / Notion-team backend is a thoughtful touch for solo-vs-agency. The honest caveats are ordinary for a young single-maintainer tool: the manual channels still need a human paste-and-click, the anti-spam/subreddit rules are only as good as the curated catalog, and idempotency/retry behaviour can't be exercised from the homepage. For an AI-assisted developer or DevRel team that wants distribution as a callable tool rather than a SaaS dashboard, it's a genuinely useful, agent-first surface.
Why STEADY
STEADY (74) for a focused, agent-native tool that solves a concrete distribution workflow with two design decisions that matter — idempotency on (content_id, channel) and a no-LLM-in-server split — clearly evidenced on the surface, MIT-licensed and npx-installable with eight documented MCP tools. Not higher because the actual publish/idempotency/retry behaviour and the anti-spam rules can't be verified from the public page, four of eight channels still require a human paste step, and it's a young single-maintainer project with adoption unproven. Not FADING because the tool surface, tool contract and honest scoping show real, thoughtful engineering.
Public-surface checklist
- PASS Homepage loads (required)
- PASS Primary value prop (required) — A genuinely agent-native content-distribution MCP: your agent writes the copy, t
- PASS Cta present
- PASS Evidence or demo — 1 screenshot(s) captured
What we saw
1 screenshot captured by the Hlido engine during the reviewed run (run-d14a0c28698f9b6a-automatelab-tech). Our own captures — not vendor marketing material.
What it does well
- Eight namespaced MCP tools (post_publish/schedule/drain/status/unpublish, channel_hints, profile_list, subreddit_list) with a clear one-envelope-plus-variants contract
- Idempotent on (content_id, channel): a re-run of the same envelope is a no-op — no double-posts or duplicate Reddit threads (as stated)
- Zero LLM calls in the server: host supplies copy/credentials/decisions, MCP supplies I/O — model- and host-agnostic
- npx -y install with one JSON block, Node 18+ only; works over stdio or SSE against Claude Code, Cursor, n8n and the plain mcp client
- Honest tiering of channels (AUTO vs MANUAL) instead of pretending every platform has a usable API; dual YAML-local / Notion-team backend
What it fails at
- Four of eight channels (Reddit, Medium, LinkedIn, Twitter) are MANUAL — a human still pastes and clicks Publish; not fully autonomous distribution
- Idempotency, retry and anti-spam/subreddit-cooldown behaviour are design claims that can't be exercised or verified from the public surface
- Young, single-maintainer project (@ratamaha); adoption, maintenance cadence and long-term platform-API resilience are unproven on the surface
- unpublish is best-effort only (Reddit is honor-system, browser channels operator-driven)
Red flags
- Idempotency and anti-double-post behaviour are the core value and are asserted, not demonstrated, on the public surface — validate against your own channels before trusting it for unattended publishing.
- Depends on multiple third-party platform APIs (some free-tier and explicitly 'hostile'/'unusable'); channel reliability is outside the tool's control.
Best for
- AI-assisted developers and DevRel teams who want cross-posting as a callable MCP tool, not a SaaS dashboard
- Agent pipelines that need idempotent publishing with per-channel canonical-URL handling for SEO
- Solo developers (YAML backend) or small agencies (Notion backend) publishing to developer-community channels
Not recommended for
- Teams wanting fully hands-off publishing to Reddit/LinkedIn/Medium/Twitter (those remain manual paste-and-click)
- Non-technical marketers who want a hosted GUI scheduler (Buffer/Hootsuite) rather than an MCP server they configure with tokens
- Anyone needing guaranteed platform-API stability — free-tier APIs (Twitter, Reddit) are explicitly called out as unreliable
Pricing & access
- ModelOpen source
- Free entry pointYes — a free tier or open-source edition is documented
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 2026-08-21.
Compared to
-
Zapier
agent-native-mcp-vs-hosted-no-code-automation
Zapier is a mature hosted automation platform with a GUI, huge app catalog and managed reliability; Content Distribution MCP is a narrow, code-first MCP server that an AI agent drives directly, with idempotency and canonical-URL handling built for developer-community channels specifically. Choose Zapier for breadth and a no-code UI, this for agent-native, MIT-licensed, tokens-you-control publishing.
Agent relevance
CLI MCP Behavioral-testable
Agentic-Commerce Readiness 63/100 · INTEGRABLE
Independent readiness for agent delegation & transaction. How it’s scored · check live
First-class MCP server (npx @ratamaha/content-distribution-mcp) exposing eight tools over stdio or SSE; any MCP host (Claude Code, Cursor, n8n, plain mcp client) drives publishing directly. No LLM calls in the server, so behaviour is deterministic and directly testable given channel credentials. A genuine agent-to-agent distribution surface.
Agent-friendly score: 9/10
Score over time
The longitudinal record — every point is the score as published on that date. Raw series.
Evidence
- MCP server publishing one piece of content to eight developer-community channels, MIT-licensed, npx @ratamaha/content-distribution-mcp — source (2026-08-21) verified
- Eight MCP tools namespaced by domain (post_publish/schedule/drain/status/unpublish, channel_hints, profile_list, subreddit_list) — source (2026-08-21) verified
- Idempotent on (content_id, channel); re-running the same envelope is a no-op — source (2026-08-21) verified
- Zero LLM calls in the server; works with any model/host over stdio or SSE (Claude Code, Cursor, n8n, mcp client) — source (2026-08-21) verified
- Three publish tiers: six AUTO native-API channels, remaining channels MANUAL draft+compose-URL — source (2026-08-21) verified
