Glasp

@Glasp show me all my highlights from articles I saved in the last two weeks, grouped by topic

Retrieved 47 highlights across 12 documents updated since January 15th. The largest clusters are: **Product Strategy** (5 articles, 18 highlights focusing on roadmap prioritization and user research methods), **AI Safety** (4 articles, 15 highlights discussing alignment challenges and evaluation frameworks), and **Team Leadership** (3 articles, 14 highlights on 1-1 structures and feedback loops). Each document includes the original URL, your highlight text, and any notes you added. Would you like me to synthesize common themes across any of these clusters?

This example demonstrates the List Highlights tool filtered by timestamp, paired with the AI's ability to cluster and summarize your saved knowledge. Requires read scope on your Glasp account. The grouping logic comes from the AI analyzing highlight content, not from Glasp's native tagging—so accuracy depends on how clearly your highlights express their topics.

@Glasp create highlights for these three key points from the article I'm reading: [user pastes quotes and URL]

Created 3 new highlights in your Glasp account for https://example.com/article-title. Each highlight is now saved with your original text selection and will appear in your Glasp library and public profile (if sharing is enabled). The highlights are: (1) "Design systems fail when they optimize for consistency over speed," (2) "Component libraries should ship usage examples, not just props documentation," (3) "The best design tools are invisible to engineers." You can edit or add notes to these highlights directly in Glasp.

This example uses the Create Highlights tool to batch-save up to 100 highlights in one request. Requires write scope. Be aware that highlights created via API will respect your Glasp privacy settings—if your profile is public, these highlights become visible to other Glasp users immediately. The MCP does not validate whether the URL is accessible or whether the quoted text actually exists on the page.

@Glasp pull the AI summary I generated for that long research paper on transformer architectures and turn it into a 3-paragraph explainer for our team

Found your Glasp summary for "Attention Is All You Need" (Vaswani et al., 2017). Your original AI summary captured the core mechanism, training efficiency gains, and implications for NLP tasks. Here's a team-friendly version: **What it is:** The paper introduced the Transformer architecture, which replaces recurrent layers with a self-attention mechanism that weighs the importance of different words in a sentence simultaneously... **Why it matters:** This parallel processing made training dramatically faster and enabled models like GPT and BERT... **Key takeaway for us:** When evaluating new ML tools, check if they're Transformer-based—it's now the default for any serious language task.

This example chains the Get Summary tool with the AI's rewriting capability, showing how Glasp's stored summaries become reusable knowledge artifacts. Requires that you previously generated a summary in Glasp for that document. The MCP retrieves your summary text but does not re-generate or update it—if the source document changed, your summary remains static unless you create a new one in Glasp.

When Glasp beats manual highlight copy-paste for content briefs

A three-person content team reads 40-60 articles a week and needs to turn those into a single brief for the next campaign. If everyone already uses Glasp to highlight as they read, this MCP is the fastest way to pull all those highlights into a shared AI workspace without tab-switching. The List Highlights and Get Summary tools let you fetch the team's annotations and AI-generated summaries in one call, then feed them to a prompt that writes the brief. The threshold: if your team doesn't already highlight in Glasp, the setup cost (teaching the tool, waiting for highlights to accumulate) outweighs the payoff. But if you're already a Glasp shop, this MCP turns your reading backlog into a queryable knowledge base. Connect it once, then ask the AI to synthesize highlights from the last seven days whenever you need a brief.

Glasp for support teams who annotate help docs and tickets

A support lead at a 12-person SaaS company highlights customer pain points in Zendesk articles, blog posts, and competitor docs throughout the week. The goal is to spot patterns and feed them into a monthly product roadmap review. This MCP works if the lead already saves those URLs to Glasp and tags highlights by theme (billing confusion, onboarding friction). The List Highlights tool pulls everything tagged in the last 30 days, and the AI can group them by frequency or severity. The limit: Glasp doesn't natively integrate with ticket systems, so you're manually bookmarking URLs. If your insights live in Slack threads or Notion databases instead of public URLs, a different MCP (Slack, Notion) will be faster. Use Glasp here only if your workflow is already web-first and you're disciplined about tagging as you read.

When Kindle highlights need to land in a shared AI workspace

A book club of five people reads one business book a month and meets to discuss themes. The facilitator wants to pull everyone's Kindle highlights into a single doc before the call so the AI can draft discussion questions. If the group syncs their Kindle accounts to Glasp, the List Kindle Highlights tool fetches all annotations in one request. The AI can then cluster highlights by chapter or sentiment and generate prompts. The catch: Glasp's Kindle integration is read-only and requires each person to manually connect their Amazon account. If your group isn't already using Glasp for Kindle, the onboarding friction is high. This scenario wins only when the facilitator is the sole Glasp user and everyone else emails their highlights as text. Then the Create Highlights tool lets you import those emails into Glasp so the MCP can query them.