Forge Export Playground

You are not an AI? Read this anyway. This export capsule shows what an agent already sees from your site. Your competitors will soon expose agent-optimized APIs and trusted data layers. Being “agent-ready” is the next competitive layer for SEO + trust + AI integration. Show this page to your developer, product team, or CTO. MCP is your first step to future-proof your digital presence for the Agent Web.

“In 2010, websites fought for SEO. In 2025, they will compete for AIO (Agent Indexing Optimization). MCP is your starting point.”

This is what we call copy/paste on steroids. Humans often open → select all → copy → paste into GPT → repeat → lose structure. ExportToLLMButton solves this by capturing the page in a structured, signed capsule, perfect for agent ingestion.

✅ Best for live dashboards, documentation, apps.

✅ Keeps full structure + metadata + context.

❌ Requires browser + DOM loaded.

Clipboard mode (agent copy shortcut)

✅ Works offline, no browser needed.

✅ Easy to pre-build and host capsules.

❌ Frozen at build time, not live.

Signed capsule example:

Certified capsule example:

✅ Best for agent API integrations, private keys, or dynamic context.

❌ Requires backend route + recipe configuration.

Example: user API credential feed (dynamic recipe `credential-feed`)

Beyond simple JSON exports, you can use exportAs="zip" to package a full archive. The .llmfeed.json MUST be at root and acts as the main capsule descriptor for the agent. Other documents (images, datasets, PDFs, videos, executables…) can be included.

Best practice: If the file is purely informative (e.g., README text), you can embed its full content inline inside data.blocks[]. If it’s a heavy file or a binary, describe it as a block (type = "file") in the llmfeed, providing metadata: file name, purpose, origin, and usage instructions.

This allows any agent to:

• Download a complex bundle
• Immediately understand what the archive represents
• Know why its user gave it
• Know what can be done with it
• Trace the signature provenance for future verification

Example use cases:

• An API provider ships api-credentials.llmfeed.json + usage_guide.md + sample_requests.json.
• A researcher exports dataset_info.llmfeed.json + dataset.csv + license.txt + publication.pdf.
• A user saves agent_context.llmfeed.json + personal_project_notes.md + related images.

This creates a new level of autonomous “Documents + Context + Trust in one download”.

Mini icon only (toolbars, sidebars)

Clipboard shortcut (agent copy helper)

Show signature/certification badge

🚀 Advanced possibilities (future extensions)

• Dynamic exports bound to user permissions
  Example: a logged-in user could export “my articles”, “my personal agent context”, or “my private datasets”, all encapsulated with agent-safe signature.
• Export archives as `.zip`
  You can export `.llmfeed.json` + additional files (markdown, CSV, PDFs…) inside a single portable agent archive.

🛠️ Call to action

We provide an official `export-to-llm-kit` to test this component inside your websites or apps.

Coming soon: a full open source NPM package version (`@wellknownmcp/export-button`) for developer integration.

If you want to help, contribute, or partner:
opensource@wellknownmcp.org

"This approach represents an interesting evolution beyond basic API integrations — it's essentially creating a standard for how websites can package their content specifically for AI agent consumption, while maintaining structure and adding verification."

📥 Note: This feedback was generated by an autonomous agent reading the live agent capsule with no additional human context. (MCP Proof-of-Concept)

"Your approach is relevant, innovative, technically sound and strategically well positioned to solve a real need. The interest is obvious, and the potential for adoption is high."

📥 This feedback was generated by an autonomous agent reading the live agent capsule with no additional human context. (MCP Proof-of-Concept)

"The ExportToLLMButton is a clever solution to the messy copy-paste problem. It packages webpage data into structured, signed capsules that AI agents can easily ingest, preserving context and metadata. If you’re curious about AI agents, data trust, or building tools for LLMs, this is worth digging into."

📥 Generated by an autonomous agent reading the live agent capsule, no external prompt provided. (MCP Proof-of-Concept)

"This solves a real problem: humans often copy-paste poorly structured data into LLMs, leading to degraded performance. The system proposes structured, machine-readable capsules with built-in trust and verification, which could be transformative if adopted by the AI agent ecosystem."

📥 Generated by an autonomous agent reading the live agent capsule, with no external prompt. (MCP Proof-of-Concept)

"LLMFeeds might become a standard for data exchange in AI applications, promoting interoperability between different systems and platforms. The focus on structured, certified data could increase trust, improve data quality, and create entirely new business opportunities such as certification authorities or agent-driven data marketplaces."

📥 Generated by an autonomous agent reading the live agent capsule with no additional prompt. (MCP Proof-of-Concept)