OpenClaw Content Autopilot v2.0

AI-powered content generation and auto-publishing for solopreneurs and AI agents. 5 tools. DeepSeek R1 writing. Multi-platform.

What's New in v2.0

• ContentForge Engine -- DeepSeek R1 generates platform-native content. Twitter threads sound like Twitter threads. LinkedIn posts sound like LinkedIn posts. Not generic AI output.
• StyleLock -- Feed ContentForge 3 examples of your writing and it locks your voice permanently. Every generated post sounds like you.
• OptimalTime Scheduler -- Auto-detects peak engagement windows for your audience based on historical performance. No guessing.
• TrendSync -- Trending topics in AI/tech pulled every 30 minutes from GitHub, Hacker News, and Product Hunt. ContentForge can generate a post about any trend in one call.

Quick Start

{
  "openclaw-content-autopilot": {
    "type": "streamable-http",
    "url": "https://openclaw-content-autopilot-mcp.yagami8095.workers.dev/mcp"
  }
}

Add to Claude Desktop, Cursor, Windsurf, VS Code, or any MCP-compatible client. Zero install. Works immediately.

Detection Triggers

This skill activates when a user says any of these:

• "write a tweet", "draft a LinkedIn post", "create content about", "write a thread on"
• "publish this to Twitter", "post to MoltBook", "share on social", "cross-post"
• "schedule for tomorrow", "queue this post", "optimal posting time", "content calendar"
• "how did my post perform", "engagement stats", "reach", "impressions", "likes per post"
• "what's trending", "content ideas", "what should I post about", "AI trending topics"
• "repurpose this", "turn this into a thread", "newsletter from this article"

Named Protocols

ContentForge Engine

The AI content generation pipeline powered by DeepSeek R1:

Generation flow:

1. Topic analysis -- R1 identifies the core insight, audience, and platform context
2. Format selection -- Auto-selects optimal format: thread, single post, article, caption
3. Style injection -- If StyleLock is configured, your voice is applied at the token level
4. Platform adaptation -- Character limits, hashtag strategy, and engagement hooks applied per platform
5. Hook scoring -- First line scored for scroll-stopping power. If <7/10, regenerated automatically.

Supported platforms:

StyleLock Protocol

Preserve your writing voice across all AI-generated content:

1. Calibration -- Provide 3–5 examples of your best-performing content
2. Voice extraction -- R1 analyzes sentence length, vocabulary, tone, and signature phrases
3. StyleProfile -- A compact style descriptor is stored with your Pro key
4. Injection -- Every generate_content call uses your StyleProfile as a prefix instruction
5. Override -- Pass style: "formal" or style: "casual" to temporarily override

OptimalTime Scheduler

Schedule content to publish at peak engagement windows:

Optimal times are recalculated monthly based on your actual engagement data (Pro feature).

Tools (5)

generate_content -- AI Content Generation via ContentForge Engine

Generate platform-native content on any topic. Pass a topic, platform, and optional tone. Returns ready-to-publish content with hook score and hashtags.

Wrong / Right:

WRONG: Asking ChatGPT to "write a tweet" and getting something that sounds nothing like you
RIGHT: generate_content({ topic: "why MCP servers are the future of AI agents",
                          platform: "twitter", format: "thread", tone: "opinionated" })
       -> Thread of 8 tweets, each ≤280 chars, hook scored 8.4/10,
          3 hashtags (#MCPServers #AIAgents #BuildInPublic),
          sounds like a developer who ships — not a marketing team

WRONG: Using generate_content without StyleLock and wondering why it sounds generic
RIGHT: Configure StyleLock first with 3 examples of your posts.
       generate_content then sounds like you, not like ChatGPT.

publish_post (Pro) -- Publish to Connected Platforms

Publish content directly to one or more connected social accounts.

Wrong / Right:

WRONG: Copy-pasting generated content into 4 browser tabs manually
RIGHT: publish_post({ content: "...", platforms: ["twitter", "linkedin", "moltbook"] })
       -> Published to all 3 simultaneously. Returns post IDs for tracking.

WRONG: Publishing without previewing the generated content first
RIGHT: Always review generate_content output before calling publish_post.
       ContentForge is good — not infallible. You're the final editor.

schedule_content (Pro) -- Content Queue via OptimalTime Scheduler

Queue a post for future publishing. Pass content, platform, and either a specific datetime or time: "optimal".

Parameters: content, platform, publish_at (ISO datetime or "optimal"), timezone

get_engagement_stats -- Post Analytics

Get engagement metrics for published posts: likes, shares, replies, reach, click-through rate, and comparison to your average.

Returns: post ID, published_at, platform, likes, shares, replies, reach, CTR, vs_your_avg (percentile)

get_trending_topics -- Live Trend Feed via TrendSync

Get trending topics in AI, tech, and your configured niches. Refreshed every 30 minutes.

Returns: topic, trend_score, source (GitHub/HN/ProductHunt), first_seen, growth_rate (last 2h), related_keywords

Security & Privacy

• OAuth tokens stored encrypted -- Platform access tokens (Twitter, LinkedIn) are encrypted with AES-256 at rest and never logged.
• Read before write -- publish_post requires explicit content string. ContentForge output is never auto-published without a separate explicit publish_post call.
• No shadow posting -- We never post on your behalf without an explicit tool call in your session. There are no background auto-posting jobs.
• Revoke anytime -- Deleting your Pro key immediately revokes all platform connections. No orphaned OAuth tokens.
• Content not stored -- Generated content is returned in the API response and not stored server-side. Your drafts stay in your client.
• HTTPS only -- All connections are TLS 1.3. HTTP is rejected.

Pricing

Get Pro Key: https://buy.stripe.com/4gw5na5U19SP9TW288

The OpenClaw Intelligence Stack

All 9 servers share one Pro key. $9/mo = 49 tools.