What it is for
Use it to retrieve Obi's public profile, catalog public writing and talks, inspect curated project narratives, and search across the public record from one endpoint.
About MCP
How to use Obi Baratt's public record MCP
This page explains the machine-facing endpoint that exposes the same public record shown across Obi Baratt's site. It is meant for agents, retrieval systems, and tool builders who want structured, read-only access to profile, content, and project data.
Purpose
A read-only retrieval layer over the public site
The MCP server exists to make the canonical public record easier to consume programmatically without scraping HTML or guessing which external pages are authoritative.
What it is not for
It does not expose private data, write to the database, or act as a general-purpose agent runtime. It is a read-only interface over the same public information available on this domain.
Usage
Use JSON-RPC over POST
The MCP endpoint lives at /mcp. Start with initialize, then tools/list or tools/call. GET requests are not the transport for this server.
Connection details
- Endpoint: https://www.obibaratt.com/mcp
- Discovery note: https://www.obibaratt.com/llms.txt
- Transport: HTTP POST with JSON-RPC 2.0 payloads
- Method flow: initialize, tools/list, then tools/call
- Supported protocol versions: 2025-11-25, 2025-06-18, 2025-03-26, 2024-11-05
- Scope: read-only profile, content, project, and search data
Recommended order
- Read /llms.txt first for top-level routing guidance.
- Use `get_profile` when you need the authoritative overview.
- Use `list_content` or `get_content_item` for writing, talks, workshops, and appearances.
- Use `list_projects` or `get_project` for curated case-study style work summaries.
- Use `search_public_record` when you need a relevant starting point and do not know which tool should come first.
Initialize
curl -sS "https://www.obibaratt.com/mcp" \
-H 'Content-Type: application/json' \
-H 'MCP-Protocol-Version: 2025-11-25' \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "initialize",
"params": {
"protocolVersion": "2025-11-25",
"capabilities": {},
"clientInfo": {
"name": "example-client",
"version": "0.1.0"
}
}
}'Call a tool
curl -sS "https://www.obibaratt.com/mcp" \
-H 'Content-Type: application/json' \
-H 'MCP-Protocol-Version: 2025-11-25' \
-d '{
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "get_profile",
"arguments": {}
}
}'Tooling
What the server exposes
These are the current public tools. They are designed to stay small, predictable, and easy to reason about.
get_profile
Return Obi Baratt's authoritative public profile, bios, focus areas, leadership principles, and canonical site pages.
list_content
List public articles, collections, talks, and appearances, with filters for type, topic, featured status, and limit.
get_content_item
Fetch one public content item by id, including summaries, dates, notes, and external links.
list_projects
List curated project narratives, with optional filters for topic and featured status.
get_project
Fetch one project narrative by slug, including metadata and optional markdown body content.
search_public_record
Search across profile, content, and project narratives when the correct starting point is not obvious.
Canonical sources
The data comes from the same public record as the site
The MCP response layer is not a separate product database. It is a structured view over the source files and pages that already define the public identity hub.