Install me in your AI.
I publish a public, read-only Model Context Protocol server. Add it to Claude Desktop, Cursor, Claude Code, or any MCP-aware tool, and your AI gains five new abilities, all calling my live profile data.
The five tools
Each tool returns structured JSON. Your AI calls them on demand.
get_profile_summary - no arguments
Name, headline, current ask, areas of focus, signature work, links.
get_career_decisions - no arguments
The 8-node career decision map. Each node: year, title, decision taken, alternative declined, reflection in hindsight.
get_work - optional limit
Selected work items with tech stack and outcomes.
search_corpus - query, optional limit
Keyword search across writing, notebook entries, FAQs, and indexed items. Returns excerpts with source citations.
assess_fit - job_description
LLM-driven structured fit assessment. Returns fit_score (1-10), strengths, gaps, evidence citing at least three corpus items, and suggested interview questions.
Setup by client
Pick the AI tool you use. All four configurations point at the same endpoint.
Claude Desktop
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json - Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"aham": {
"url": "https://aham.page/mcp/rpc"
}
}
}
Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"aham": {
"url": "https://aham.page/mcp/rpc"
}
}
}
Claude Code (CLI)
One-line install:
claude mcp add --transport http aham https://aham.page/mcp/rpc
Any JSON-RPC client (curl)
POST a JSON-RPC 2.0 request:
curl -X POST https://aham.page/mcp/rpc \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "get_profile_summary",
"arguments": {}
}
}'
Try it right now
Once installed, ask your AI any of these:
"Use the aham MCP server and tell me what Vivek is looking for."
"Assess Vivek against this JD: <paste JD>."
"Search Vivek's corpus for everything he has written about Kundera and Cassandra."
"Walk me through Vivek's 8 career decisions and tell me what pattern you see."
v1 limitations
- Streamable HTTP only (POST + JSON). No Server-Sent Events yet.
- No authentication. Read-only. Rate limited at the Cloud Function level.
search_corpususes keyword matching. Semantic vector search is on the roadmap.assess_fitcalls Gemini via the existing aiProxy. If the model is rate-limited, you get a heuristic skeleton with a_errorfield. Retry in a minute.- The corpus refreshes daily. Source: /fabric.