AT Protocol MCP Server

Appearance

recommend_content ​

Recommend posts based on user interests and engagement history. Analyzes your timeline, engagement patterns, and network to suggest relevant content.

Authentication ​

Required - This tool requires authentication to analyze your engagement history and generate personalized recommendations.

Parameters ​

ParameterTypeRequiredDefaultDescription
maxResultsnumberNo20Maximum number of recommended posts to return. Must be between 1 and 100.
minLikesnumberNo5Minimum like count for recommended posts. Must be 0 or greater.
maxAgenumberNo24Maximum age of posts in hours. Must be between 1 and 168 (7 days).
topicsstring[]No-Optional array of topics to filter recommendations.
excludeRepostsbooleanNofalseWhether to exclude reposted content from recommendations.

Response ​

typescript
{
  success: boolean;
  recommendations: Array<{
    uri: string;
    cid: string;
    author: {
      did: string;
      handle: string;
      displayName?: string;
      avatar?: string;
    };
    text: string;
    likeCount: number;
    replyCount: number;
    repostCount: number;
    indexedAt: string;
    recommendationScore: number;
    recommendationReasons: string[];
    topics?: string[];
  }>;
  insights: string[];
}

Examples ​

Get Personalized Recommendations ​

json
{
  "maxResults": 20,
  "minLikes": 5,
  "maxAge": 24
}

Find High-Quality Recent Content ​

json
{
  "maxResults": 30,
  "minLikes": 20,
  "maxAge": 12,
  "excludeReposts": true
}

Topic-Specific Recommendations ​

json
{
  "maxResults": 25,
  "topics": ["technology", "AI", "programming"],
  "minLikes": 10
}

Discover Older Content ​

json
{
  "maxResults": 50,
  "maxAge": 168,
  "minLikes": 15
}

Error Handling ​

Common errors:

  • AuthenticationRequired: Must be authenticated to use this tool
  • InvalidRequest: Invalid parameters
  • InsufficientData: Not enough engagement history to generate recommendations
  • RateLimitExceeded: Too many requests in a short period

Best Practices ​

  1. Regular Discovery: Run daily to discover new content
  2. Adjust Filters: Experiment with different minLikes and maxAge values
  3. Use Topics: Filter by topics when looking for specific content
  4. Review Reasons: Understand why content is recommended
  5. Engage with Recommendations: Like, reply, or repost to improve future recommendations
  6. Exclude Reposts: Set excludeReposts: true for original content only
  7. Vary Time Windows: Try different maxAge values to discover both fresh and evergreen content

Recommendation Score ​

The recommendation score (0-100) is calculated based on:

  • Engagement: Likes, reposts, and replies
  • Author Preference: Posts from authors you frequently engage with
  • Topic Relevance: Match with your interests and specified topics
  • Recency: Newer posts score higher (within the time window)
  • Network Overlap: Posts from your network or similar users

Recommendation Reasons ​

Common reasons include:

  • "High engagement (X likes)"
  • "From an author you frequently engage with"
  • "Matches your topic interests"
  • "Popular in your network"
  • "Similar to posts you've liked"

Rate Limiting ​

This tool is subject to AT Protocol API rate limits:

  • 3,000 requests per hour for authenticated users
  • May require multiple API calls to analyze timeline and engagement

See Also ​